Worked on performance optimization for the deepmodeling/abacus-develop repository, focusing on enhancing plane wave data processing for large datasets. Applied C++ and OpenMP to optimize FFT transform copy routines in the PW_Basis module, introducing cache blocking and SIMD techniques to improve cache utilization and processing speed. Parallelized the count_pw_st function using OpenMP, refining the parallelization strategy with collapse adjustments to achieve better multi-core utilization and stability. The work targeted computational efficiency, leveraging parallel programming and performance optimization skills to deliver a feature that accelerates data processing workflows without introducing new bugs, demonstrating depth in algorithmic and low-level systems engineering.