Investigating Nvidia GPU Architecture Trends via Microbenchmarks

  • Lingqi Zhang
  • , Ryan Barton
  • , Peng Chen
  • , Xiao Wang
  • , Toshio Endo
  • , Satoshi Matsuoka
  • , Mohamed Wahib

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

Abstract

In recent years, GPUs have become crucial tools in high performance computing. To understand the evolution of these devices, we conducted a study analyzing hardware trends in GPUs. This information provides valuable insights into GPU programming and emerging trends in this field. We mainly focus on three features: the floating-point operations per second (FLOPS) of floating-point units (FPUs), device memory access, and machine balance.

Original languageEnglish
Title of host publicationProceedings - 2024 IEEE International Conference on Cluster Computing Workshops, CLUSTER Workshops 2024
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages174-175
Number of pages2
ISBN (Electronic)9798350383454
DOIs
StatePublished - 2024
Event2024 IEEE International Conference on Cluster Computing Workshops, CLUSTER Workshops 2024 - Kobe, Japan
Duration: Sep 24 2024Sep 27 2024

Publication series

NameProceedings - 2024 IEEE International Conference on Cluster Computing Workshops, CLUSTER Workshops 2024

Conference

Conference2024 IEEE International Conference on Cluster Computing Workshops, CLUSTER Workshops 2024
Country/TerritoryJapan
CityKobe
Period09/24/2409/27/24

Keywords

  • GPUs
  • Microbenchmarks

Fingerprint

Dive into the research topics of 'Investigating Nvidia GPU Architecture Trends via Microbenchmarks'. Together they form a unique fingerprint.

Cite this