Add real implementation #1

Current implementation uses doubles to calculate PI:

https://github.com/Fewlaps/superpi/blob/37f0f138147a386c2cd21451f96463e5c618ec5f/src/main/java/com/fewlaps/superpi/PiCalculator.java#L11-L14

however, as you probably know double has precision of 15-17 decimal numbers (see wiki). So your code actually is not able to calculate more digits, and therefore it doesn't actually implement superpi algorithm.

I propose to implement it correctly.