import numpy as np
N, M = map(int, input().split())
A = [sum(map(int, input().split())) for _ in range(N)]
dp1 = A[:]
dp2 = [0]*N
if N>1:
  dp2[1] = dp1[0]-A[1]
for i in range(2,N):
  dp1[i] = max([dp1[i],dp1[i-1],dp2[i-1]+A[i]])
  dp2[i] = max(dp1[i-1]-A[i],dp2[i-1])
print(max(dp1[-1],dp2[-1]))