IMI: In-memory Multi-job Inference Acceleration for Large Language Models

Author:

Gao Bin1ORCID,Wang Zhehui2ORCID,He Zhuomin3ORCID,Luo Tao2ORCID,Wong Weng-Fai1ORCID,Zhou Zhi4ORCID

Affiliation:

1. National University of Singapore, Singapore

2. Institute of High-Performance Computing, A*STAR, Singapore

3. Shanghai Jiao Tong University, China

4. Sun Yat-sen University, China

Publisher

ACM

Reference36 articles.

1. Aayush Ankit, Izzat El Hajj, Sai Rahul Chalamalasetti, Sapan Agarwal, Matthew Marinella, Martin Foltin, John Paul Strachan, Dejan Milojicic, Wen-Mei Hwu, and Kaushik Roy. 2020. Panther: A programmable architecture for neural network training harnessing energy-efficient reram. IEEE Trans. Comput. (2020).

2. Suyu Ge, Yunan Zhang, Liyuan Liu, Minjia Zhang, Jiawei Han, and Jianfeng Gao. 2023. Model tells you what to discard: Adaptive kv cache compression for llms. arXiv preprint arXiv:2310.01801 (2023).

3. Transformer in transformer;Han Kai;Advances in Neural Information Processing Systems,2021

4. Mingxuan He, Choungki Song, Ilkon Kim, Chunseok Jeong, Seho Kim, Il Park, Mithuna Thottethodi, and TN Vijaykumar. 2020. Newton: A DRAM-maker’s accelerator-in-memory (AiM) architecture for machine learning. In IEEE/ACM MICRO.

5. [5] Intel. 2024. https://www.intel.com/content/www/us/en/developer/articles/technical/memory-performance-in-a-nutshell.html.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3