Accelerating Deep Learning Using Interconnect-Aware UCX Communication for MPI Collectives-Reference-Cited by-同舟云学术

登录注册会员服务联系我们

Accelerating Deep Learning Using Interconnect-Aware UCX Communication for MPI Collectives

Published:2022-03-01 Issue:2 Volume:42 Page:68-76
ISSN:0272-1732
Container-title:IEEE Micro
language:
Short-container-title:IEEE Micro

Author:

Temucin Yltan Hassan¹^ORCID,Sojoodi Amir Hossein¹^ORCID,Alizadeh Pedram¹,Kitor Benjamin¹^ORCID,Afsahi Ahmad¹

Affiliation:

1. Queen’s University, Kingston, ON, Canada

Funder

Natural Sciences and Engineering Research Council of Canada

Compute Canada

Western Canada Research Grid

SciNet

Compute Canada Calcul Canada

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Subject

Electrical and Electronic Engineering,Hardware and Architecture,Software

Link

http://xplorestaging.ieee.org/ielx7/40/9743780/09705571.pdf?arnumber=9705571

Reference11 articles.

1. Efficient Multi-Path NVLink/PCIe-Aware UCX based Collective Communication for Deep Learning

2. Communication Profiling and Characterization of Deep Learning Workloads on Clusters with High-Performance Interconnects

3. CUDA Kernel Based Collective Reduction Operations on Large-scale GPU Clusters

4. Design considerations for GPU-aware collective communications in MPI

5. Scalable Distributed DNN Training using TensorFlow and CUDA-Aware MPI: Characterization, Designs, and Performance Evaluation

Cited by 9 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. ROCm-Aware Leader-based Designs for MPI Neighbourhood Collectives;ISC High Performance 2024 Research Paper Proceedings (39th International Conference);2024-05

2. Analysis and prediction of performance variability in large-scale computing systems;The Journal of Supercomputing;2024-03-28

3. Enhancing Intra-Node GPU-to-GPU Performance in MPI+UCX through Multi-Path Communication;Proceedings of the 3rd International Workshop on Extreme Heterogeneity Solutions;2024-03-02

4. Designing Solutions for High-Performance Communications Software Design in Network Applications;2024 International Conference on Optimization Computing and Wireless Communication (ICOCWC);2024-01-29

5. Exploiting copy engines for intra-node MPI collective communication;The Journal of Supercomputing;2023-05-11

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