perf the compute of check_atom_stru and add openmp

[A-006]

Linked Issue

Fix #3720

What's changed?

• To optimize the computation, we adopted an addition-based approach instead of using division and square root operations. The cell numbers are pre-computed before executing the main function to enhance performance.
• Thanks to @dzzz2001, we integrated OpenMP to leverage parallel computing in our algorithm. This implementation demonstrates strong scaling, indicating significant performance improvements with increased computational resources.

Here is the time table of the perf and openmp