All-Reduce Communication is a collective communication operation in distributed computing where each process contributes a local data element, all elements are globally reduced using a specified associative and commutative reduction operator, and the resulting global reduction is delivered back to every participating process, distinguishing it from other reduction topologies by its bidirectional aggregation-dissemination pattern and all-process-equality semantic.
4
Publications
121
Citations
8
Authors
4
Institutions
Publications per year
2017–2026
2
Publications
4
Kylix: A Sparse Allreduce for Commodity Clusters
Huasha Zhao, John Canny · 2014 · 36 citations
DeAR: Accelerating Distributed Deep Learning with Fine-Grained All-Reduce Pipelining
Lin Zhang, Shaohuai Shi, Xiaowen Chu et al. · 2023 · 17 citations
1–5 of 4
Authors
8
- SP
1
h-index
- JC
1
h-index
- XC
1
h-index
- QA
1
h-index
- HZ
1
h-index
1–5 of 8
Institutions
4
Hong Kong Baptist University
Hong Kong, Hong Kong
1 h-index
1 publications
34 citations
Purdue University West Lafayette
West Lafayette, United States
1 h-index
3 publications
42 citations
University of California, Berkeley
Berkeley, United States
1 h-index
2 publications
72 citations
Hong Kong University of Science and Technology
Hong Kong, Hong Kong
1 h-index
2 publications
68 citations
Harbin Institute of Technology
Harbin, China
1 publications
17 citations
Shenzhen Technology University
Shenzhen, China
1 publications
17 citations
Imperial College London
London, United Kingdom
1 publications
11 citations
1–7 of 4
Venues
- Edinburgh Research Explorer
Repository · 1 publications · 11 citations
- Journal of Optical Communications and Networking
JOCN · 1 publications · 12 citations
1–2