Improving Password Generation Algorithm with Parallellism: Comparative Performance Study

Passwords are crucial for online security, serving as the primary defense for critical accounts and sensitive data. However, evolving cyber threats challenge password resilience, necessitating robust solutions. This paper investigates PassGAN, a deep-learning approach that surpasses traditional methods by leveraging machine learning algorithms for password generation. The study optimizes PassGAN’s performance through parallel computing, addressing time-intensive complexity. Experiments show a significant improvement in PassGAN’s efficiency with multiprocessing. Using four processing elements, it generates 500,000 passwords in 1024 s, demonstrating a 3.1x speedup. This parallelization empowers PassGAN to efficiently produce diverse passwords, enhancing its overall effectiveness in password security. This research underscores PassGAN’s potential, augmented by parallel computing, as a formidable tool against cyber threats, emphasizing its ability to enhance secure password generation.