Virtual Channel aware Scheduling for Real Time Data-Flows on Network on-Chip

Autor: Abou El Hassan Benyamina, Mohammed Amine Meghabber, Lakhdar Loukil, Richard Olejnik, Abdelkader Aroui
Přispěvatelé: Université des Sciences et Technologies d'Oran, Département d'Informatique [Oran], Université des sciences et de la Technologie d'Oran Mohamed Boudiaf [Oran] (USTO MB), Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 (CRIStAL), Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS)
Rok vydání: 2019
Předmět:
Zdroj: Scalable Computing: Practice and Experience
Scalable Computing: Practice and Experience, 2019, 20, pp.495-509. ⟨10.12694/scpe.v20i3.1423⟩
ISSN: 1895-1767
DOI: 10.12694/scpe.v20i3.1423
Popis: International audience; The increasing complexity of real-time applications presents a challenge to researchers and software designers. The tasks of these applications usually exchange many data-flows and often need to satisfy real-time constraints. Although the Network on-Chip (NoC) paradigm offers an underlying communication infrastructure that gives more hardware resources, tasks and data-flows cannot be performed before their deadlines. In recent works, preemptive wormhole switching with fixed priority has been introduced to meet real-time constraints of real-time applications. However, it suffers some bottleneck such as hardware requirement where none of these works takes account of the number of implemented Virtual Channels on the router. To alleviate this problem, we propose a novel scheduler for soft real-time data-flows application that takes into account the lack on resource in routers in term of Virtual Channels. Experimental results obtained on a benchmark of synthetic and soft real applications have shown the efficiency of our approach in term of real-time constraints satisfaction for data-flow traffics and hardware requirements.
Databáze: OpenAIRE