Optimizing GPU Cache Policies for MI Workloads∗

Johnathan Alsop, Matthew D. Sinclair, Srikant Bharadwaj, Alexandru Dutu, Anthony Gutierrez, Onur Kayiran, Michael Lebeane, Brandon Potter, Sooraj Puthoor, Xianwei Zhang, Tsung Tai Yeh, Bradford M. Beckmann

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

Abstract

In recent years, machine intelligence (MI) applications have emerged as a major driver for the computing industry. Optimizing these workloads is important, but complicated. As memory demands grow and data movement overheads increasingly limit performance, determining the best GPU caching policy to use for a diverse range of MI workloads represents one important challenge. To study this, we evaluate 17 MI applications and characterize their behavior using a range of GPU caching strategies. In our evaluations, we find that the choice of caching policy in GPU caches involves multiple performance trade-offs and interactions, and there is no one-size-fits-all GPU caching policy for MI workloads. Based on detailed simulation results, we motivate and evaluate a set of cache optimizations that consistently match the performance of the best static GPU caching policies.

Original languageEnglish
Title of host publicationProceedings of the 2019 IEEE International Symposium on Workload Characterization, IISWC 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages243-248
Number of pages6
ISBN (Electronic)9781728140452
DOIs
StatePublished - Nov 2019
Event15th IEEE International Symposium on Workload Characterization, IISWC 2019 - Orlando, United States
Duration: 3 Nov 20195 Nov 2019

Publication series

NameProceedings of the 2019 IEEE International Symposium on Workload Characterization, IISWC 2019

Conference

Conference15th IEEE International Symposium on Workload Characterization, IISWC 2019
CountryUnited States
CityOrlando
Period3/11/195/11/19

Keywords

  • execution-driven simulation
  • GPU caching
  • machine intelligence
  • machine learning

Fingerprint Dive into the research topics of 'Optimizing GPU Cache Policies for MI Workloads∗'. Together they form a unique fingerprint.

  • Cite this

    Alsop, J., Sinclair, M. D., Bharadwaj, S., Dutu, A., Gutierrez, A., Kayiran, O., Lebeane, M., Potter, B., Puthoor, S., Zhang, X., Yeh, T. T., & Beckmann, B. M. (2019). Optimizing GPU Cache Policies for MI Workloads∗. In Proceedings of the 2019 IEEE International Symposium on Workload Characterization, IISWC 2019 (pp. 243-248). [9041977] (Proceedings of the 2019 IEEE International Symposium on Workload Characterization, IISWC 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IISWC47752.2019.9041977