Parallelizing highly complex engine management systems. (23rd April 2017)

• Record Type: Journal Article
• Authors: Kienberger, Julian; Schmidhuber, Stefan; Saad, Christian; Kuntz, Stefan; Bauer, Bernhard
• Other Names: Lengauer Christian guestEditor.; Bougé Luc guestEditor.; Trystram Denis guestEditor.; Balaji Pavan guestEditor.; Leung Kai‐Cheung guestEditor.
• Is Part Of: Concurrency and computation. Volume 29:Number 15(2017)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2017-04-23
• Subjects: multi‐core -- model‐based design -- data dependency analysis -- timing simulation -- design spaceexploration -- semi‐automated parallelization; Parallel processing (Electronic computers) -- Periodicals; Parallel computers -- Periodicals; 004.35
• DOI: 10.1002/cpe.4115
• Languages: English
• ISSNs: 1532-0626

Abstract

Summary: The automotive industry seeks to include more and more features in its vehicles. For this purpose, the necessary policy shift towards multi‐core technology is in full swing. To eventually exploit the extra processing power, there is much additional effort needed for coping with the tremendously increased complexity. This is largely due to the elaborate parallelization process that spans a vast search space. Consequently, there is a strong need for innovative methods and appropriate tools for the migration of legacy single‐core software. We use the results of a data dependency analysis performed on AUTOSAR system descriptions to determine advantageous partitions as well as initial task‐to‐core mappings. Afterwards, the extracted information serves as input for the simulation within a multi‐core timing tool suite. Here, the initial solution is evaluated with respect to proper scheduling and metrics like cross‐core communication rates, communication latencies, or core load distribution. A subsequent optimization process improves the initial solution and enables a comparative assessment. To demonstrate the benefit, we substantially expand a previous case study by applying our approach to two complex engine management systems and by showing the advantages compared to a parallelization process without preceding dependency analysis and initial partition/mapping suggestions.