I offer a small program for counting primes by quadratic interval. The program simply counts how many primes occur from one perfect square to the next. (It uses trial division under 10^{12} and MillerRabin above that. However, it fails due to the limitation of the language's log function not far above that  said language I shall not reveal. Otherwise, it would continue into the e20s over the course of weeks, months, or years....) The PNT count is the basic version, using an unrefined primecounting function  that is, N divided by the natural logarithm of N. The formula for applying this "expected" number to the interval between one perfect square and the next is constructed thus:
((n+1)^{2}/log((n+1)^{2}))  (n^{2}/log(n^{2}))
(Thanks to members of the MathForum for setting me straight about this.)
Quadratic Interval Prime Counter
Download (18KB)
Generated values are: Perfect Square, Interval Count,
Prime Count, Prime Ratio, Log Count, Log Ratio, Deviation
The results up to 100 million are viewable and downloadable here.
A graph comparing the crude primecounting function and the real count looks like this  showing the real count exceeding the expected count for most intervals  and deviating by larger amounts with larger intervals.
This graph does NOT show the cumulative prime count: real or expected. It shows the prime count from one perfect square to the next. This is something quite different than you will most likely see at other math sites where the primecounting function is explained and graphed as a cumulative total, with the line showing a steplike pattern. This graphs show the total number of primes for each successive quadratic interval  and it demonstrates that the number of primes is increasing absolutely for successive intervals.
The deviation between the real and logderived count for each interval looks like this:
Close up, it looks like this...
