Proceedings of the Korea Information Processing Society Conference (한국정보처리학회:학술대회논문집)

  • 2024.05a
  • /
  • Pages.31-34
  • /
  • 2024
  • /
  • 2005-0011(pISSN)
  • /
  • 2671-7298(eISSN)
Korea Information Processing Society (한국정보처리학회)
권호 표지

Accelerating and analyzing the Recommendation System using Processing-in-Memory

Processing-in-Memory 를 이용한 추천시스템 가속화 및 분석

  • Jung-uk Hong (Dept. of Electrical and Computer Engineering, Seoul National University ) ;
  • Jin-ho Lee (Dept. of Electrical and Computer Engineering, Seoul National University)
  • 홍정욱 (서울대학교 전기정보공학부 ) ;
  • 이진호 (서울대학교 전기정보공학부)
  • Published : 2024.05.23
Download PDF
⟨ Previous Next ⟩

Abstract

추천 시스템(Recommendation System)은 인터넷 쇼핑몰, 넷플릭스, SNS 등 여러 분야에서 유저에게 맞는 타겟 광고를 추천하는 시스템을 말한다. 추천 시스템을 가속하기 위해서는 추천 시스템 모델에서 불규칙적이고 잦은 데이터 이동으로 인해 병목현상을 일으키는 임베딩 레이어를 타겟하는 것이 중요하다고 알려져 있다. 이 논문에서는 데이터 이동이 잦은 어플리케이션에 효과적인 Processing-in-Memory 를 이용하여 추천 시스템을 가속하고 분석한다.

Keywords

References

  1. M. Naumov et al., "Deep learning recommendation model for personalization and recommendation systems", arXiv preprint arXiv: 1906.00091, 2019. 
  2. U. Gupta et al., "The Architectural Implications of Facebook's DNN-Based Personalized Recommendation", in IEEE International Symposium on High Performance Computer Architecture, 2 2020, pp 488-501. 
  3. J. Gomez-Luna et al., '"Benchmarking a new paradigm: Experimental analysis and characterization of a real processing-in-memory system", IEEE Access, vol. 10, pp. 52 565-52 608, 2022. 
  4. H. Ye et al., "GRACE: A Scalable Graph-Based Approach to Accelerating Recommendation Model Inference", in Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Vancouver BC Canada: ACM, 3 2023, Volume 3, pp 282-301. 
  5. L. Ke et al., "RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing", in 2020 ACM/IEEE 47th Annual International Symposium on Computer Architecture (ISCA), Valencia, Spain: IEEE, 5 2020, pp 790-803. 
  6. D. Mudigere et al., "Software-hardware co-design for fast and scalable training of deep learning recommendation models" in Proceedings of the 49th Annual International Symposium on Computer Architecture, 2022, pp 993-1011. 
  7. E. Chan et al., "Collective communication: theory, practice, and experience.", Concurrency and computation: Practice and Experience, vol. 10, no. 13, pp. 1749-1783, 2007. 
  8. "Criteolabs Kaggle display advertising challenge dataset." https://labs.criteo.com/2014/02/download-kaggle-display-advertising-challenge-dataset/ 
  9. "UPMEM SDK." https://sdk.upmem.com/ 
  10. S. U. Noh et al., "PID-Comm: A Fast and Flexible Collective Communication Framework for Commodity Processing-in-DIMM Devices.", in 2024 ACM/IEEE 51th Annual International Symposium on Computer Architecture, 2024.