Thèse soutenue

Génération automatique d'architectures multiprocesseurs hétérogènes : aspects logiciels et matériels

FR  |  
EN
Auteur / Autrice : Youenn Corre
Direction : Jean-Philippe DiguetLoïc Lagadec
Type : Thèse de doctorat
Discipline(s) : STIC
Date : Soutenance en 2013
Etablissement(s) : Lorient
Ecole(s) doctorale(s) : École doctorale Santé, information-communication et mathématiques, matière (Brest, Finistère)
Partenaire(s) de recherche : autre partenaire : Université européenne de Bretagne (2007-2016)

Résumé

FR  |  
EN

Embedded systems evolution has led to the emergence of H-MPSoCs which provide a way to respect the cost and performance constraints inherent to embedded systems. However they also make the task of designing and programming such systems a long and arduous process. It is thus necessary to develop tools that will free designers from architectural and programming details, so that they can focus on the tasks where they can bring added-value. The objective is thus to automatize the tasks that burden the design of H-MPSoC, in particular on FPGA, by providing a higher-level of abstraction following a method that brings together HLS and hardware/software co-design beyond the existing solutions which are whether incomplete or unfit. The presented work introduces a design framework relying on the automation of tedious tasks and allowing designers to express their expertise where they want to. For this, we rely on an architecture model defined with a high-level formalism independent from implementation details, providing a solution to the lack of multiprocessor architecture in FPGAs. This specification model also allows designers to provide design constraints in accordance with their level of expertise or involvement. The DSE is implemented as a scalable algorithm relying on fast and accurate estimation techniques. A method for the exploration of hardware accelerators based on HLS to provide fast cost estimations is introduced. The use of MDE methods enables portability and reuse by generating the final design implementation. The framework is validated through two case studies: an MJPEG video decoder and a face detection application.