On the performance of scalable fast optical switch based recursive network architecture for HPC applications

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

Abstract

We investigate a novel scalable low-latency HPC network for blade-to-blade interconnection exploiting multiple-parallel recursive architecture with fast optical switches under realistic HPC application traffics. Compared with Leaf-Spine HPC interconnecting 4096-blade, results show that HFOS4 reduces up to 75.9% latency, 64.2% cost and 30.9% power consumption.

Original languageEnglish
Title of host publication45th European Conference on Optical Communication, ECOC 2019
PublisherInstitution of Engineering and Technology (IET)
ISBN (Electronic)9781839530074, 9781839530661, 9781839530883, 9781839530890, 9781839531071, 9781839531088, 9781839531255, 9781839531705, 9781839531859
Publication statusPublished - 2019
Event45th European Conference on Optical Communication, ECOC 2019 - Dublin, Ireland
Duration: 22 Sep 201926 Sep 2019
https://www.ecoc2019.org/

Publication series

NameIET Conference Publications
NumberCP765

Conference

Conference45th European Conference on Optical Communication, ECOC 2019
Abbreviated titleECOC 2019
CountryIreland
CityDublin
Period22/09/1926/09/19
Internet address

Keywords

  • High performance computing
  • Network architecture
  • Optical switches
  • System performance

Fingerprint Dive into the research topics of 'On the performance of scalable fast optical switch based recursive network architecture for HPC applications'. Together they form a unique fingerprint.

Cite this