Efficient Modeling of Embedded Systems Using Computer-Aided Recoding

The design of embedded computing systems, such as video-enabled mobile phones and reliable medical devices, faces a serious productivity gap due to the increasing complexity of their hardware and software, as well as hard design constraints, including strict timing, multi-core functionality, low power and price, and short time-to-market. One solution to address this problem is the modeling at higher levels of abstraction. Much like the quality of an architectural blue-print determines the quality of the resulting building, the model of an embedded system is the key to its successful implementation. However, writing executable system models is challenging, error-prone, and very time-consuming. In this talk, we outline a novel modeling technique called computer-aided recoding which automates various steps in the process of specifying and modeling embedded systems by use of advanced computer-aided design (CAD) techniques. Our automation relieves the system designer from complex code analysis and tedious code writing tasks, allowing uninterrupted focus on the actual system and its design space exploration. Using a designer-controlled approach with automated source code transformations, computer-aided recoding can derive an executable parallel system model directly from available sequential reference code. As a result, system modeling is streamlined, enabling a shorter design time and higher productivity, as well as quality improvements in the end design.