Loading [MathJax]/extensions/TeX/extpfeil.js
Automatic paral ellizati on of Simulink models for multicore embedded systems development | IEEE Conference Publication | IEEE Xplore

Automatic paral ellizati on of Simulink models for multicore embedded systems development


Abstract:

The embedded software systems are first designed and validated by high level models such as MATLAB/Simulink models. The Simulink is a model-based design tool, widely used...Show More

Abstract:

The embedded software systems are first designed and validated by high level models such as MATLAB/Simulink models. The Simulink is a model-based design tool, widely used in the field of control and signal processing systems for ranging from algorithm development to code generation for embedded systems. However, with the increased use of multicore technology, designers are now required to consider this technology that the software might deployed on. For instance, it is very important to parallelize software in order to exploit the inherent performance of multicore processors. This paper proposes a method to parallelize models developed on Simulink. Under the parallelization process, the proposed method uses the flattening technique to extract task-parallelism embedded in the model's hierarchy and then measure their parallel execution time when running on multicore architecture. On the experiment of applying the method for a MIMO-OFDM application model, the execution time of the parallelized model could be reduced successfully to 50.75% on 2-cores and to 25.07% on a 4-core processor, compared with that running on single core.
Date of Conference: 20-22 October 2017
Date Added to IEEE Xplore: 29 March 2018
ISBN Information:
Conference Location: Gafsa, Tunisia

References

References is not available for this document.