An ad-hoc parallelism approach for accelerating compute-intensive mobile workloads

Recently, the processing power of mobile devices has increased along with the breadth of applications that can run on them. Running these programs on mobile devices presents a significant barrier due to their processing complexity. In this work, a system for effective task execution in the context of mobile ad hoc parallelism is proposed as a solution to this problem. A new resource allocation scheme for heterogeneous mobile ad hoc cloud (Het-MAHC) is proposed by leveraging information about each mobile device's processing capacity and link lifetime. The proposed approach uses this information in the selection and partitioning process and enables parallel computing and transmission to reduce execution time and improve performance. Using Network Simulator 3, the new scheme is assessed, and its performance is evaluated against the state-of-the-art method. The proposed system is also integrated into the Android OS, and its performance was evaluated on real smartphone devices. Results show substantial performance improvements over existing approaches.