Primality test. My second contribution.

This article describes a better result than that described in the article "Primality test. My contribution".

With this review I want to show how I have performed an additional 10 million checks; of which 8 with negative results.

The reference is still Fermat's little theorem.

However, I changed the base from 2 to 3; so in this article I start from (3^a-3)/a.

I consider (3^a-3)/a as the comparison of two sequences which I call (a) and (b); it is evident that the numbers present in the sequence (b) are a function (b=3^a-3) of the corresponding numbers present in the sequence (a).

As described in the previous article I have again worked on the reason for the growth of the numbers belonging to the sequence (b) starting by removing the dependence on the corresponding numbers of the sequence (a).

According to the previous verifications I made, I found a beginning and a reason for growth for the numbers belonging to the sequence (b) such that the result of b/a is always an integer if the number of the sequence (a) is a prime number and on the contrary b / a is always a fractional number if the number of the sequence (a) is a composite number.