During the Industry Partner Program at the HiPEAC CSW, Dr.-Ing. Timo Stripf presents on Efficient Embedded Programming of Heterogeneous Multicore Platforms.
Latest embedded systems are increasingly equipped with high-performance multicore processors as well as FPGAs and GPUs. This is essential, to keep up with high performance and energy requirements of applications. The parallelization of software is currently done manually in most industrial companies. This manual process is very complex, expensive, time consuming and challenging for embedded software developers. For these reasons, many industrial companies even void multicore deployment completely. In this presentation, a solution for automated C code development and parallelization from MATLAB®, Simulink and Scilab for homogenous and heterogeneous embedded multicore systems is introduced. Tests revealed that the presented solution could decrease the effort for parallelization up to 80% and that the C code development process is simplified significantly. This enables companies using this technology to profit from important competitive advantages such as time and cost lead.