International Symposium on Future of Computer Technology 2018 (ISFCT2018)
Automatic parallelization for the embedded industry
Improvements in sequential software optimization are slowing down. As a result, software parallelization is becoming more and more popular to improve computing performance. Originating from academic high-performance computing projects, today, parallelization techniques can even augment performance of embedded applications such as gasoline engine management systems.
This presentation introduces how the benefits of parallel execution can be transferred to industrial settings in the embedded domain. One challenge is to achieve both highest reliability and a high degree of analysis automation at the same time. Furthermore, integrating automatic parallelization into existing industry workflows requires a new class of tools and functionality. Finally, examples from real-world settings highlight how customers are enabled to overcome these issues by employing the OSCARTech® Compiler.
Click here for the slide
Christoph Schumacher is an application engineer at Oscar Technology Corporation, Tokyo. He also serves as Guest Assistant Professor at the Advanced Multicore Research Institute of Waseda University, Tokyo. In both of his roles, he takes on the challenge to apply automatic parallelization technology to real-world software.
Christoph obtained a doctoral degree from RWTH Aachen University, Germany, in 2015, where he investigated the construction of parallel and distributed SystemC simulators. His current interests include processes and tools to lower the bar for the integration of automatic parallelization technologies.