Programming the adapteva epiphany 64-core network-on-chip coprocessor

Anish Varghese, Bob Edwards, Gaurav Mitra, Alistair P. Rendell

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    30 Citations (Scopus)

    Abstract

    With energy efficiency and power consumption being the primary impediment in the path to exascale systems, low-power high performance embedded systems are of increasing interest. The Parallella System-on-module (SoM) created by Adapteva combines the Epiphany-IV 64-core coprocessor with a host ARM processor housed in a Zynq System-on-chip. The Epiphany integrates low-power RISC cores on a 2D mesh network and promises up to 70 GFLOPS/Watt of processing efficiency. However, with just 32 KB of memory per eCore for storing both data and code, and only low level inter-core communication support, programming the Epiphany system presents several challenges. In this paper we evaluate the performance of the Epiphany system for a variety of basic compute and communication operations. Guided by this data we explore various strategies for implementing stencil based application codes on the Epiphany system. With future systems expected to house 4096 eCores, the merits of the Epiphany architecture as a path to exascale is compared to other competing power efficient systems.

    Original languageEnglish
    Title of host publicationProceedings - IEEE 28th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
    PublisherIEEE Computer Society
    Pages984-992
    Number of pages9
    ISBN (Electronic)9780769552088
    DOIs
    Publication statusPublished - 27 Nov 2014
    Event28th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014 - Phoenix, United States
    Duration: 19 May 201423 May 2014

    Publication series

    NameProceedings - IEEE 28th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014

    Conference

    Conference28th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
    Country/TerritoryUnited States
    CityPhoenix
    Period19/05/1423/05/14

    Fingerprint

    Dive into the research topics of 'Programming the adapteva epiphany 64-core network-on-chip coprocessor'. Together they form a unique fingerprint.

    Cite this