Modélisation flux de données et optimisation pour architecture multi-coeurs de motifs répétitifs

FACE AU DEFI QUE REPRESENTE LA PROGRAMMATION DES ARCHITECTURES MULTI-CŒURS/PROCESSEURS, IL EST DEVENU NECESSAIRE DE PROPOSER AUX DEVELOPPEURS DES OUTILS ADAPTES PERMETTANT D ABSTRAIRE LES NOTIONS INHERENTES AU PARALLELISME ET FACILITANT LE PORTAGE D UNE APPLICATION SUR DIFFERENTES ARCHITECTURES. LA METHODOLOGIE AAA (ADEQUATION ALGORITHME ARCHITECTURE) PROPOSE AU DEVELOPPEUR D AUTOMATISER LES ETAPES DE PARTITIONNEMENT, ORDONNANCEMENT A PARTIR D UNE DESCRIPTION HAUT NIVEAU DE L APPLICATION ET DE L ARCHITECTURE. CETTE METHODOLOGIE PERMET DONC LE PROTOTYPAGE RAPIDE D UNE APPLICATION SUR DIFFERENTES ARCHITECTURES AVEC UN MINIMUM D EFFORT ET UN RESULTAT APPROCHANT L OPTIMAL. LES APPORTS DE CETTE THESE SE SITUENT A LA FOIS AU NIVEAU DU MODELE DE SPECIFICATION ET DE SES OPTIMISATIONS RELATIVES AU CONTEXTE DES ARCHITECTURES PARALLELES.SINCE APPLICATIONS SUCH AS VIDEO CODING/DECODING OR DIGITAL COMMUNICATIONS WITH ADVANCED FEATURES ARE BECOMING MORE COMPLEX, THE NEED FOR COMPUTATIONAL POWER IS RAPIDLY INCREASING. IN ORDER TO SATISFY SOFTWARE REQUIREMENTS, THE USE OF PARALLEL ARCHITECTURE IS A COMMON ANSWER. TO REDUCE THE SOFTWARE DEVELOPMENT EFFORT FOR SUCH ARCHITECTURES, IT IS NECESSARY TO PROVIDE THE PROGRAMMER WITH EFFICIENT TOOLS CAPABLE OF AUTOMATICALLY SOLVING COMMUNICATIONS AND SOFTWARE PARTITIONING/SCHEDULING CONCERNS. THE ALGORITHM ARCHITECTURE MATCHING METHODOLOGY HELPS THE PROGRAMMER BY PROVIDING AUTOMATIC TRANSFORMATION, PARTITIONING AND SCHEDULING OF AN APPLICATION FOR A GIVEN ARCHITECTURE THIS METHODOLOGY RELIES ON AN APPLICATION MODEL THAT ALLOW TO EXTRACT THE AVAILABLE PARALLELISM. THE CONTRIBUTIONS OF THIS THESIS ARE TACKLES BOTH THE PROBLEM OF THE MODEL AND THE ASSOCIATED OPTIMIZATION FOR PARALLELISM EXTRACTION.RENNES-INSA (352382210) / SudocSudocFranceF