NeutronHeter: Optimizing Distributed Graph Neural Network Training for Heterogeneous Clusters

Summary: NeutronHeter: top-down multi-level mapping on a tree-like resource graph (hierarchical clustering of compute and bandwidth) to solve the multi-constrained multi-way GNN placement problem in heterogeneous clusters. Plus adaptive communication migration via selective vertex replication to bypass low-bandwidth links, reducing asymmetric communication hotspots and yielding 1.06–33.05x speedups over SOTA. (summarized by gpt-5-mini on Feb 11 2026)

Paper ID: 7522
Venue: SIGMOD
Year: 2026
Pagerank: 5.093636e-05
Overall Rank: 10,323 | 29.18%
DOI: 10.1145/3749175

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Chunyu Cao (Northeastern University)
2. Xin Ai (Northeastern University)
3. Qiange Wang (Northeastern University)
4. Yanfeng Zhang (Northeastern University)
5. Zhenbo Fu (Northeastern University)
6. Hao Yuan (Northeastern University)
7. Mingyi Cao (Northeastern University)
8. Chaoyi Chen (Northeastern University)
9. Yingyou Wen (Neusoft AI Magic Technology Research)
10. Yu Gu (Northeastern University)
11. Ge Yu (Northeastern University)

BibTeX Citation

@inproceedings{cao_sigmod26,
        title = {{NeutronHeter: Optimizing Distributed Graph Neural Network Training for Heterogeneous Clusters}},
        author = {Cao, Chunyu and Ai, Xin and Wang, Qiange and Zhang, Yanfeng and Fu, Zhenbo and Yuan, Hao and Cao, Mingyi and Chen, Chaoyi and Wen, Yingyou and Gu, Yu and Yu, Ge},
        series = {{SIGMOD} '26},
        booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
        publisher = {Association for Computing Machinery},
        doi = {10.1145/3749175},
        url = {https://dl.acm.org/doi/10.1145/3749175},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 2 of 2 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
10,241	FastGNAS: Accelerating and Scaling Graph Neural Architecture Search on Multi-GPUs via Ring-Based Model Migration	2026	SIGMOD	5.093636e-05
10,596	NeutronCloud: Resource-Aware Distributed GNN Training in Fluctuating Cloud Environments	2026	VLDB	5.093636e-05

Outgoing Citations (Sorted by Pagerank)

Showing 15 of 15 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
223	AliGraph: A Comprehensive Graph Neural Network Platform	2019	VLDB	0.00024182473
1,132	SANCUS: Staleness-Aware Communication-Avoiding Full-Graph Decentralized Training in Large-Scale Graph Neural Networks	2022	VLDB	0.00012041292
1,863	ByteGNN: Efficient Graph Neural Network Training at Large Scale	2022	VLDB	9.5950349e-05
2,162	Heterogeneity-aware Distributed Parameter Servers	2017	SIGMOD	9.0581831e-05
2,640	Scalable and Efficient Full-Graph GNN Training for Large Graphs	2023	SIGMOD	8.3074486e-05
2,668	DUCATI: A Dual-Cache Training System for Graph Neural Networks on Giant Graphs with the GPU	2023	SIGMOD	8.2750247e-05
2,695	NeutronStar: Distributed GNN Training with Hybrid Dependency Management	2022	SIGMOD	8.2468134e-05
5,079	NeutronOrch: Rethinking Sample-based GNN Training under CPU-GPU Heterogeneous Environments	2024	VLDB	6.3717342e-05
6,588	Lotan: Bridging the Gap between GNNs and Scalable Graph Analytics Engines	2023	VLDB	5.833338e-05
6,630	Efficient Training of Graph Neural Networks on Large Graphs	2024	VLDB	5.8192133e-05
6,803	OUTRE: An OUT-of-core De-REdundancy GNN Training Framework for Massive Graphs within A Single Machine	2024	VLDB	5.7684339e-05
6,806	HongTu: Scalable Full-Graph GNN Training on Multiple GPUs	2023	SIGMOD	5.7673207e-05
6,857	RAGraph: A Region-Aware Framework for Geo-Distributed Graph Processing	2024	VLDB	5.7525746e-05
7,087	DAHA: Accelerating GNN Training with Data and Hardware Aware Execution Planning	2024	VLDB	5.7069166e-05
7,353	XGNN: Boosting Multi-GPU GNN Training via Global GNN Memory Store	2024	VLDB	5.6352669e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	9,729	Scalable Graph Convolutional Network Training on Distributed-Memory Systems	2023	VLDB
2	1,863	ByteGNN: Efficient Graph Neural Network Training at Large Scale	2022	VLDB
3	4,878	NeutronStream: A Dynamic GNN Training Framework with Sliding Window for Graph Streams	2024	VLDB
4	2,640	Scalable and Efficient Full-Graph GNN Training for Large Graphs	2023	SIGMOD
5	10,890	Heta: Distributed Training of Heterogeneous Graph Neural Networks	2025	VLDB
6	5,079	NeutronOrch: Rethinking Sample-based GNN Training under CPU-GPU Heterogeneous Environments	2024	VLDB
7	2,695	NeutronStar: Distributed GNN Training with Hybrid Dependency Management	2022	SIGMOD
8	10,835	NeutronTask: Scalable and Efficient Multi-GPU GNN Training with Task Parallelism	2025	VLDB
9	9,546	NeutronTP: Load-Balanced Distributed Full-Graph GNN Training with Tensor Parallelism	2025	VLDB
10	10,596	NeutronCloud: Resource-Aware Distributed GNN Training in Fluctuating Cloud Environments	2026	VLDB