Parallel DNN Training With Uniform Local Batch in Heterogeneous Environments
- 1 December 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 37 (12) , 2908-2920
Abstract
No abstract availableFunding Information
- Natural Science Foundation of Jilin Province (20230101062JC)
- Research Project on Undergraduate Education and Teaching Reform at Jilin University
- National Natural Science Foundation of China (62272190)
- Innovation Funding of ICT
- CAS (E461050)
This publication has 22 references indexed in Scilit:
- DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learningNature, 2025
- ArrayPipe: Introducing Job-Array Pipeline Parallelism for High Throughput Model Exploration2025
- Alleviating straggler impacts for data parallel deep learning with hybrid parameter updateFuture Generation Computer Systems, 2025
- AntDT: A Self-Adaptive Distributed Training Framework for Leader and Straggler Nodes2024
- ExplSched: Maximizing Deep Learning Cluster Efficiency for Exploratory Jobs2023
- iGniter: Interference-Aware GPU Resource Provisioning for Predictable DNN Inference in the CloudIEEE Transactions on Parallel and Distributed Systems, 2022
- Varuna2022
- Semi-dynamic load balancing2020
- Falcon: Addressing Stragglers in Heterogeneous Parameter Server Via Multiple ParallelismIEEE Transactions on Computers, 2020
- Heterogeneity and Interference-Aware Virtual Machine Provisioning for Predictable Performance in the CloudIEEE Transactions on Computers, 2015