Glasswing: Accelerating MapReduce on multi-core and many-core clusters

Ismail Elhelw, R.F.H. Hofman, H.E. Bal

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

Abstract

The impact and significance of parallel computing techniques is continuously increasing given the current trend of incorporating more cores in new processor designs. However, many Big Data systems fail to exploit the abundant computational power of multicore CPUs and GPUs to their full potential. We present Glasswing, a scalable MapReduce framework that employs a configurable mixture of coarse- and fine-grained parallelism to achieve high performance on multi-core CPUs and GPUs. We experimentally evaluated the performance of five MapReduce applications and show that Glasswing outperforms Hadoop on a 64-node multi-core CPU cluster by a factor between 1.8 and 4, and by a factor from 20 to 30 on a 16-node GPU cluster. Copyright © 2014 ACM.
Original languageEnglish
Title of host publicationHPDC 2014 - Proceedings of the 23rd International Symposium on High-Performance Parallel and Distributed Computing
PublisherAssociation for Computer Machinery
Pages295-298
Number of pages4
ISBN (Print)9781450327480
DOIs
Publication statusPublished - 2014
Event23rd ACM Symposium on High-Performance Parallel and Distributed Computing, HPDC 2014 - Vancouver, BC, Canada
Duration: 23 Jun 201427 Jun 2014

Conference

Conference23rd ACM Symposium on High-Performance Parallel and Distributed Computing, HPDC 2014
CountryCanada
CityVancouver, BC
Period23/06/1427/06/14

Keywords

  • Heterogeneous
  • MapReduce
  • Opencl
  • Scalability

Fingerprint Dive into the research topics of 'Glasswing: Accelerating MapReduce on multi-core and many-core clusters'. Together they form a unique fingerprint.

Cite this