Specializing Coherence, Consistency, and Push/Pull for GPU Graph Analytics

Giordano Salvador, Wesley H. Darvin, Muhammad Huzaifa, Johnathan Alsop, Matthew D. Sinclair, Sarita V. Adve

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This work explores the interaction of three communication-centric design dimensions for graph workloads on emerging integrated CPU-GPU systems: update propagation with and without fine-grained synchronization (push vs. pull), emerging coherence protocols (GPU vs. DeNovo coherence), and software-centric consistency models (DRF0, DRF1, and DRFrlx). We show that these dimensions are inter-dependent and the best design depends on the graph algorithm and input. We develop a model to predict this best design, motivating flexible and hardware-software co-designed GPU memory systems.

Original languageEnglish (US)
Title of host publicationProceedings - 2020 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages123-125
Number of pages3
ISBN (Electronic)9781728147987
DOIs
StatePublished - Aug 2020
Event2020 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2020 - Boston, United States
Duration: Aug 23 2020Aug 25 2020

Publication series

NameProceedings - 2020 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2020

Conference

Conference2020 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2020
Country/TerritoryUnited States
CityBoston
Period8/23/208/25/20

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Hardware and Architecture
  • Safety, Risk, Reliability and Quality

Cite this