Optimize logrow + fix benchmark folder

[JernKunpittaya]

This issue was already mentioned in benchmark folder itself, see more info in that folder's README, but basically

'logrow' can be chosen arbitrarily between min(log_2(number of rows), log_2(lookup range)) to max(log_2(number of rows), log_2(lookup range)). The trade off is that if we choose lower logrow value , we will face with more "columns for non-linearity table".

For now, based on our experiment, the best value to select logrow is to select the lowest logrow in the range such that ceiling(lookup_range/2^logrow) <=5

Hence,

1. We should incorporate this optimization into zkstats itself natively
2. Due to some changes in zkstats flow from merging PR#34, we should update our benchmark folder accordingly