Moment: Co-optimizing Physical Communication Topology and Data Placement for Multi-GPU Out-of-core GNN Training

作者：Zuocheng Shi, Jie Sun, Z.H. Song, Mo Sun, Yang Xiao, Fei Wu, Zeke Wang · 年份：2025 · DOI：10.1145/3712285.3759788 · 被引用次数：3 · 研究领域：Graph Theory and Algorithms、Advanced Graph Neural Networks、Advanced Neural Network Applications

Graph Neural Networks (GNNs) are widely employed in applications like recommendation systems, social network analysis, and fraud detection, but training large-scale GNNs is challenging due to its memory limitations. Existing systems face a trade-off between throughput and monetary cost: Distributed systems require expensive memory scaling, while single-machine out-of-core systems are limited by GPU/PCIe throughput. To this end, we propose Moment, a physical communication topology and data placement co-optimizer to enable high-throughput and low-cost GNN training in a single multi-GPU machine. Moment addresses communication contention and GPU load imbalance issues by modeling the physical topology as capacity-constrained directed graphs and formulating communication scheduling as a max-flow problem. It also introduces a data-distribution-aware knapsack algorithm for optimized data placement. Experimental results show that Moment outperforms out-of-core systems by up to 6.51 × and distributed systems by up to 3.02 ×, with only 50% monetary cost.