MGI: A Communication Framework for Data Processing in Massive GPU Infrastructures
Summary: MGI abstracts recurring inter-GPU communication optimizations behind a unified interface. A resource-annotated graph controller plans transfers, while a scalable device/link-aware data plane executes them efficiently across heterogeneous GPU infrastructures. (summarized by gpt-5.6-luna on Aug 17 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Di Wu (University of Toronto)
- 2. Hongshi Tan (National University of Singapore)
- 3. Hanzhang Yang (University of Toronto)
- 4. Bingsheng He (National University of Singapore)
- 5. Qizhen Zhang (University of Toronto)
BibTeX Citation
@article{wu_vldb26,
title = {{MGI: A Communication Framework for Data Processing in Massive GPU Infrastructures}},
author = {Wu, Di and Tan, Hongshi and Yang, Hanzhang and He, Bingsheng and Zhang, Qizhen},
journal = {PVLDB},
series = {{VLDB} '26},
volume = {19},
number = {9},
pages = {2262--2275},
doi = {10.14778/3819518.3819549},
url = {https://doi.org/10.14778/3819518.3819549},
year = {2026}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 24 of 24 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 3,752 | G3: When Graph Neural Networks Meet Parallel Graph Processing Systems on GPUs | 2020 | VLDB |
| 2 | 4,480 | GPU Database Systems Characterization and Optimization | 2024 | VLDB |
| 3 | 2,284 | Pump Up the Volume: Processing Large Data on GPUs with Fast Interconnects | 2020 | SIGMOD |
| 4 | 3,309 | A Distributed Multi-GPU System for Fast Graph Processing | 2018 | VLDB |
| 5 | 10,345 | Data Movement-Aware GPU Sharing for Data-Intensive Systems | 2026 | CIDR |
| 6 | 5,521 | EMOGI: Efficient Memory-access for Out-of-memory Graph-traversal in GPUs | 2021 | VLDB |
| 7 | 6,516 | GPU-accelerated data management under the test of time | 2020 | CIDR |
| 8 | 9,957 | Towards Sufficient GPU-accelerated Dynamic Graph Management: Survey and Experiment | 2025 | VLDB |
| 9 | 1,224 | Large Graph Convolutional Network Training with GPU-Oriented Data Communication Architecture | 2021 | VLDB |
| 10 | 3,053 | MG-Join: A Scalable Join for Massively Parallel Multi-GPU Architectures | 2021 | SIGMOD |