1 Introduction
Multi-cores are becoming mainstream platforms for real-time embedded systems to meet the rapidly increasing computation requirements and low power consumption [1], [2], [3]. To fully utilize the capacity of multi-cores, embedded software must be parallelized. OpenMP [4] is the de facto parallel programming framework for shared memory architectures, which can greatly simplify the development of parallel software by providing various mechanisms to explore the explicit and implicit software parallelism. OpenMP is not only widely used in high-performance and general computing, but also drawing increasing attention in the embedded and real-time computing domain [5], [6], [7], [8], [9], [10].