Software performance simulation strategies for high-level embedded system design

As most embedded applications are realized in software, software performance estimation is a very important issue in embedded system design. In the last decades, instruction set simulators (ISSs) have become an essential part of an embedded software design process. However, ISSs are either slow or very difficult to develop. With the advent of multiprocessor systems and their ever-increasing complexity, the software simulation strategy based on ISSs is no longer efficient enough for exploring the large design space of multiprocessor systems in early design phases. Motivated by the limitations of ISSs, a lot of recent research activities focused on software simulation strategies based on native execution. In this article, we first introduce some existing software performance simulation strategies as well as our own approach for source level simulation, called SciSim, and provide a discussion about their benefits and limitations. The main contribution of this article is to introduce a new software performance simulation approach, called iSciSim (intermediate Source code instrumentation based Simulation), which achieves high estimation accuracy, high simulation speed and low implementation complexity. All these advantages make iSciSim well-suited for system level design. To show the benefits of the proposed approach, we present a quantitative comparison between iSciSim and the other discussed techniques, using a set of benchmarks.