Model-Based Automotive Partitioning And Mapping For Embedded Multicore Systems

Robert H ̈ottger, Lukas Krawczyk, Burkhard Igel

Zenodo (CERN European Organization for Nuclear Research) · 2015 · 26 citations · 10 references

DOIFull text

Open access

Concepts

Abstract

This paper introduces novel approaches to partitioning<br> and mapping in terms of model-based embedded multicore system<br> engineering and further discusses benefits, industrial relevance and<br> features in common with existing approaches. In order to assess<br> and evaluate results, both approaches have been applied to a real<br> industrial application as well as to various prototypical demonstrative<br> applications, that have been developed and implemented for<br> different purposes. Evaluations show, that such applications improve<br> significantly according to performance, energy efficiency, meeting<br> timing constraints and covering maintaining issues by using<br> the AMALTHEA platform and the implemented approaches.<br> Furthermore, the model-based design provides an open, expandable,<br> platform independent and scalable exchange format between<br> OEMs, suppliers and developers on different levels. Our proposed<br> mechanisms provide meaningful multicore system utilization since<br> load balancing by means of partitioning and mapping is effectively<br> performed with regard to the modeled systems including hardware,<br> software, operating system, scheduling, constraints, configuration and<br> more data.

References

10