Network Requirements for Distributed Machine Learning Training in the Cloud
Name
Salamy-jsalamy-SM-EECS-2022-thesis.pdf
Description
Thesis PDF
Size
6.94 MB
Format
Adobe PDF
Checksum (MD5)
6b2ef3fcd9d326c3c883b8c2198cd57f
Author(s)
Salamy, James
Advisor(s)
Ghobadi, Manya
Date Issued
February 2022
Publisher
Massachusetts Institute of Technology
Abstract
In this thesis, I characterize the impact of network bandwidth on distributed machine learning training. I test four popular machine learning models (ResNet, DenseNet, VGG, and BERT) on an Nvidia A-100 cluster to determine the impact of bursty and non-bursty cross traffic (such as web-search traffic and long-lived flows) on the iteration time and throughput of distributed training. By varying the cross traffic load, I measure the impact of network congestion on training iteration times. I observe that with heavy web-search cross traffic (80% of link capacity), on average training iteration time is increased by up to 4 to 8×, for ResNet and BERT models, respectively. Further, I establish that the ring-all reduce communication collective is negatively impacted by network congestion even if the congestion is only affecting part of the ring. I also develop empirical models for the behavior of machine learning training in the presence of each type of cross traffic deployed. These results provide the motivation for developing novel congestion control protocols that are tailored for distributed training environments.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright MIT
Persistent DSpace Link
