Predictable Parallel Programming using PRET-C

Embedded system designers are always faced with the challenge of achieving a good price-performance trade-off. Recently, the advent of affordable multicore processors has offered designers the choice of high-performance architectures. However, efficient and predictable programming of multicore systems still poses considerable challenges. This paper offers a novel solution to this problem by answering two nagging questions. Firstly we address the question of efficient parallel programming using C, a language of choice in the embedded domain. Secondly, we address the issue of timing predictability, a significant concern in most embedded systems. In doing so, we propose a novel technique of parallel programming by extending C with macros and a synchronous execution semantics. We also demonstrate using extensive experimentation that the proposed approach is both efficient, scalable, while also enabling predictable execution. To our knowledge, this is the first approach for predictable parallel programming using synchronous C.