Password recovery parallelisation and acceleration technology for autonomous and controllable platforms

With the widespread use of encryption in data storage and transmission, efficient password recovery is critical for legitimate access restoration and digital forensics.Traditional CPU-based solutions are inefficient for modern cryptographic algorithms such as AES-256 and SHA-512 due to limited serial computing capability, while GPU-based acceleration suffers from memory bandwidth and latency bottlenecks under complex control logic.To address these issues, this paper proposes a high-performance password recovery system for autonomous and controllable platforms.An intelligent decryption heterogeneous acceleration architecture (IDHAA) is designed to improve resource utilisation and coordination efficiency through fine-grained task decomposition and dynamic scheduling across heterogeneous computing units.Furthermore, a heuristic dynamic optimisation search (HDOS) algorithm is introduced to reduce blind traversal in large password spaces by adaptively optimising search strategies based on structural features and feedback information.Experimental results demonstrate significant improvements in recovery efficiency, success rate and system scalability.