Optimizing makespan and resource utilization for multi-DNN training in GPU cluster

标题
Optimizing makespan and resource utilization for multi-DNN training in GPU cluster
作者
关键词
Deep neural network (DNN) training, Ring-Allreduce, Job scheduling, Resource allocation, Linear scaling rule (LSR), GPU cluster
出版商
Elsevier BV
发表日期
2021-06-24
DOI
10.1016/j.future.2021.06.021

向作者/读者发起求助以获取更多资源

Reprint

联系作者

Add your recorded webinar

Do you already have a recorded webinar? Grow your audience and get more views by easily listing your recording on Peeref.

Upload Now

Become a Peeref-certified reviewer

The Peeref Institute provides free reviewer training that teaches the core competencies of the academic peer review process.

Get Started