Optimizing DNN Compilation for Distributed Training With Joint OP and Tensor Fusion
Author:
Affiliation:
1. Computer Science, University of Hong Kong, Hong Kong, Hong Kong
2. Alibaba Cloud Intelligent, Alibaba Group, Hangzhou, Zhejiang, China
3. Compute Arch, NVIDIA Corp, Beijing, China
Funder
Alibaba Group
Hong Kong RGC contracts HKU
Publisher
Institute of Electrical and Electronics Engineers (IEEE)
Subject
Computational Theory and Mathematics,Hardware and Architecture,Signal Processing
Link
http://xplorestaging.ieee.org/ielx7/71/9790018/09866830.pdf?arnumber=9866830
Reference59 articles.
1. Simple embedding for link prediction in knowledge graphs;kazemi,2018
2. Link prediction based on graph neural networks;zhang,2018
3. Beyond data and model parallelism for deep neural networks;jia,2018
Cited by 4 articles. 订阅此论文施引文献 订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献
1. Adaptive partitioning and efficient scheduling for distributed DNN training in heterogeneous IoT environment;Computer Communications;2024-02
2. BladeDISC: Optimizing Dynamic Shape Machine Learning Workloads via Compiler Approach;Proceedings of the ACM on Management of Data;2023-11-13
3. RECom: A Compiler Approach to Accelerating Recommendation Model Inference with Massive Embedding Columns;Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 4;2023-03-25
4. A Comparative Study of Neural Network Compilers on ARMv8 Architecture;Architecture of Computing Systems;2023
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