I am trying to estimate the upper bound of the time-complexity of this solution by ksun48. Currently, I am assuming the upper bound on the number of divisors of N to be based on this codeforces blog. For simplicity, if we assume the bound on no. of divisors to be exactly N^1 / 3, the upper bound can be calculated as N^1 / 3 + N^1 / 9 + N^1 / 27 + ....

How can we find a tighter upper bound for this solution?