Database Paper Browser

Back to papers

NeutronTP: Load-Balanced Distributed Full-Graph GNN Training with Tensor Parallelism

Summary: NeutronTP uses tensor parallelism for GNNs—partitioning node features (not graph) to remove cross-worker vertex dependencies and achieve load balance. A decoupled NN/aggregation framework plus memory-efficient, comm/comp-overlapping scheduling enables multi-GPU training and 1.29–8.72× speedups. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID
13832
Venue
VLDB
Year
2025
Pagerank
4.3441378e-05
Overall Rank
9,395 | 34.65%
DOI
10.14778/3705829.3705837

Incoming Non-self Citations Over Time

Authors

Incoming Citations (Sorted by Pagerank)

Showing 3 of 3 citing papers.

Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 8 of 8 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