Service interruption on Monday 11 July from 12:30 to 13:00: all the sites of the CCSD (HAL, EpiSciences, SciencesConf, AureHAL) will be inaccessible (network hardware connection).
Skip to Main content Skip to Navigation
Conference papers

Wormhole networks properties and their use for optimizing worst case delay analysis of many-cores

Abstract : Many-core architectures are promising hardware to design hard real-time systems as they are based on simpler and thus more predictable processors than multi-core systems. However, the worst-case behavior of the Network-on-Chip (NoC) for both inter-core and core to external memories or peripherals communications must be established. Several NoCs targeting hard real-time systems, made of specific hardware extensions, have been designed. However, none of these extensions are currently available in commercially existing NoC-based many-core architectures, that instead rely on wormhole switching with round-robin arbitration. In this paper, we thus demonstrate three properties of such NoC-based wormhole networks to identify worst-case scenarios and reduce the pessimism when modeling flows in contentions. We then describe and evaluate an algorithm to compute Worst-Case Traversal Time (WCTT) of flows that uses these properties. In particular, our results show that the pessimism can be reduced up to 50% compared to current state-of-the-art real-time packet schedulability analysis.
Document type :
Conference papers
Complete list of metadata

https://hal-cea.archives-ouvertes.fr/cea-01836859
Contributor : Léna Le Roy Connect in order to contact the contributor
Submitted on : Thursday, July 12, 2018 - 3:05:26 PM
Last modification on : Monday, July 4, 2022 - 8:59:34 AM

Identifiers

Citation

L. Abdallah, M. Jan, Jérôme Ermont, Christian Fraboul. Wormhole networks properties and their use for optimizing worst case delay analysis of many-cores. 10th IEEE International Symposium on Industrial Embedded Systems (SIES), Jun 2015, Siegen, Germany. pp.59-68, ⟨10.1109/SIES.2015.7185041⟩. ⟨cea-01836859⟩

Share

Metrics

Record views

88