Optimize LinearSieve with segmented sieves

[byeongkeunahn]

Currently, LinearSieve uses a linear sieving algorithm but it is not very performant. It achieves ~600 ms up to 10**7 on Baekjoon Online Judge.

A cache-aware sieve algorithm called segmented sieves can dramatically improve the sieving performance. It is described here: https://github.com/kimwalisch/primesieve/wiki/Segmented-sieve-of-Eratosthenes

The above project is licensed under BSD-2. Hence, if the implementation is to be directly translated and adapted, it is possible, and we would need to add corresponding license acknowledgements.