The obvious and naive method described above is O(sqrt(N)). For N ~= 2 ^ 127, that is about 2 ^ 64. / The Lucas-Lehmer method described in the article is better (how much better is an exercise for the reader).
You are assuming division itself is an O(1) operation. However, it also scales with the size of the number. So more correct would be to say that this naive method is O(sqrt(N) log(N) log(log(N))).
