HopGNN：通过以特征为中心的模型迁移提升分布式 GNN 训练效率
HopGNN: Boosting Distributed GNN Training Efficiency via Feature-Centric Model Migration
AI总结：
针对分布式 GNN 训练中大规模顶点特征传输造成的通信瓶颈，提出以特征为中心的 LeapGNN，通过微图训练、特征预聚集和微图合并，相较 P3 最高加速 4.2 倍。
摘要：
图神经网络（GNN）的分布式训练已成为处理大规模图的一项关键技术。当前主流 GNN 框架以模型为中心，需要将海量图顶点特征传输到 GNN 模型，从而造成严重的通信瓶颈。认识到模型规模通常显著小于特征规模，我们提出 LeapGNN，这是一个以特征为中心的框架，通过将 GNN 模型带到顶点特征来反转这一范式。为使其真正有效，我们首先提出一种基于微图的训练策略，利用具有更优局部性的精炼结构训练模型，以减少远程特征获取。随后，我们设计了一种特征预聚集方法，将多次获取操作合并为一次，以消除冗余特征传输。最后，我们采用基于微图的合并方法，为每个 worker 调整微图数量，以最小化内核切换和同步开销。实验结果表明，与最先进方法 P3 相比，LeapGNN 实现了最高 4.2 倍的性能加速。
英文摘要：
Distributed training of graph neural networks (GNNs) has become a crucial technique for processing large graphs. Prevalent GNN frameworks are model-centric, necessitating the transfer of massive graph vertex features to GNN models, which leads to a significant communication bottleneck. Recognizing that the model size is often significantly smaller than the feature size, we propose LeapGNN, a feature-centric framework that reverses this paradigm by bringing GNN models to vertex features. To make it truly effective, we first propose a micrograph-based training strategy that trains the model using a refined structure with superior locality to reduce remote feature retrieval. Then, we devise a feature pre-gathering approach that merges multiple fetch operations into a single one to eliminate redundant feature transmissions. Finally, we employ a micrograph-based merging method that adjusts the number of micrographs for each worker to minimize kernel switches and synchronization overhead. Our experimental results demonstrate that LeapGNN achieves a performance speedup of up to 4.2x compared to the state-of-the-art method, namely P3.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
