package com.telerikacademy;

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;

public class McDuck {
    public static void main(String[] args) throws IOException {
        BufferedReader br = new BufferedReader(new InputStreamReader(System.in));
        //N and M
        String line = br.readLine();
        String[] dimensions = line.split(" ");
        int N = Integer.parseInt(dimensions[0]);
        int M = Integer.parseInt(dimensions[1]);

        // matrix NxM
        int[][] matrix = new int[N][M];

        //Where is Scrooge McDuck now
        int curX = -1;
        int curY = -1;

        //Filling the matrix with M-стойности на всеки от N-те реда
        for (int i = 0; i < N; i++) {
            String lines = br.readLine();
            String[] nLines = lines.split(" ");
            for (int j = 0; j < M; j++) {
                matrix[i][j] = Integer.parseInt(nLines[j]);
                //Finding McDuck here
                if (matrix[i][j] == 0) {
                    curX = i;
                    curY = j;
                }
            }
        }
        int coins = 0;
        while (bestMove(curX, curY, N, M, matrix) != -1 && !(curX - 1 <= 0 && curX + 1 >= 0 && curY - 1 <= 0 && curY + 1 >= 0)) {
                switch (bestMove(curX, curY, N, M, matrix)) {
                    case 4:
                        System.out.println("case 4:");
                        if (matrix[curX][curY]-- >= 0) {
                            matrix[curX][curY]--;
                            curY--;
                            coins++;
                        }
                        curY--;
                        break;
                    case 3:
                        System.out.println("case 3:");
                        if (matrix[curX][curY]-- >= 0) {
                            matrix[curX][curY]--;
                            curY++;
                            coins++;
                        }
                        curY++;
                        break;
                    case 2:
                        System.out.println("case 2:");
                        if (matrix[curX][curY]-- >= 0) {
                            matrix[curX][curY]--;
                            curX--;
                            coins++;
                        }
                        curX--;
                        break;
                    case 1:
                        System.out.println("case 1:");
                        if (matrix[curX][curY]-- >= 0) {
                            matrix[curX][curY]--;
                            curX++;
                            coins++;
                        }
                        curX++;
                        break;
                    default:
                        System.out.println("case default:");
                        System.out.println(coins);
                }
            }
        System.out.println(coins);

    }

    private static int bestMove(int curI, int curJ, int rows, int cols, int[][] matrix) {
        int direction = -1;
        System.out.println("We made a move");
        if (curJ - 1 >= 0 && matrix[curI][curJ - 1] == mostCoins(curI, curJ, matrix)) {
            System.out.println("We made a move on left");
            direction = 4;
        } else if (curJ + 1 < cols && matrix[curI][curJ + 1] == mostCoins(curI, curJ, matrix)) {
            System.out.println("We made a move on right");
            direction = 3;
        } else if (curI - 1 >= 0 && matrix[curI - 1][curJ] == mostCoins(curI, curJ, matrix)) {
            System.out.println("We made a move on up");
            direction = 2;
        } else if (curI + 1 < rows && matrix[curI + 1][curJ] == mostCoins(curI, curJ, matrix)) {
            System.out.println("We made a move on down");
            direction = 1;
        } else {
            System.out.println("We made a move out");
        }
        return direction;
    }

    private static int mostCoins(int i, int j, int[][] matrix) {
        System.out.println("We used most Coins");
        return Math.max(matrix[i][j - 1], Math.max(matrix[i][j + 1], Math.max(matrix[i - 1][j], matrix[i + 1][j])));
    }

}