/ev-store-dlrm

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

License: GPL v3 Platform

  _______     ______  _                 
 | ____\ \   / / ___|| |_ ___  _ __ ___ 
 |  _|  \ \ / /\___ \| __/ _ \| '__/ _ \
 | |___  \ V /  ___) | || (_) | | |  __/
 |_____|  \_/  |____/ \__\___/|_|  \___| -- Groupability-aware caching systems for DRS

This repository contains the implementation code for paper:
EVSTORE: Storage and Caching Capabilities for Scaling Embedding Tables in Deep Recommendation Systems

Contact Information

Maintainer: Daniar H. Kurniawan, Email: daniar@uchicago.edu

Feel free to contact Daniar for any suggestions/feedback, bug reports, or general discussions.

Please consider citing our EVStore paper at ASPLOS 2023 if you use EVStore. The bib entry is

@InProceedings{Daniar-EVStore, 
Author = {Daniar H. Kurniawan and Ruipu Wang and Kahfi S. Zulkifli and Fandi A. Wiranata and John Bent and Ymir Vigfusson and Haryadi S. Gunawi},
Title = "EVSTORE: Storage and Caching Capabilities for Scaling
Embedding Tables in Deep Recommendation Systems",
Booktitle =  {Proceedings of the 28th International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)},
Address = {Vancouver, Canada},
Month =  {MARCH},
Year =  {2023}
}

Run EVStore

Please follow the experiments detailed in Experiments.md.

Acknowledgement

The DLRM code in this repository is based on Facebook DLRM. The cache benchmark repository is based on Cache2k and Cacheus.