SNI-GNN：借助SmartNIC的全图GNN训练，采用网络内嵌入预测
SNI-GNN: SmartNIC-Assisted Full-Graph GNN Training with In-Network Embedding Prediction
浏览论文内容
中文总结 AI 辅助
SNI-GNN是借助SmartNIC的全图GNN训练系统，通过网络内嵌入预测减少通信，在保持精度的同时实现21%-45%通信量降低，1.3-3.6倍加速，可高效扩展至16个GPU。
中文摘要 AI 辅助
全图GNN训练精度高，但在多服务器集群上扩展性差，原因是节点间嵌入交换量大且不规则。本文提出SNI-GNN，一种借助SmartNIC的全图训练系统，通过在网络内预测远程嵌入减少通信量同时保持精度。SNI-GNN在SmartNIC上部署轻量线性趋势预测器，用于优化缓存的历史嵌入，结合基于重要性的边界节点采样策略，以及带中间结果复用的异步DPU-GPU数据流水线。我们给出误差和收敛界，证明在有界二阶动态下预测器偏差可被控制，且在不精确梯度下实现标准非凸收敛。在NVIDIA BlueField-3上实现的SNI-GNN与最先进的全图系统集成，通信量减少21%至45%，相较于BNS-GCN实现1.3至3.6倍的端到端加速，相较于基线SANCUS实现最高1.29倍加速，精度损失≤0.01，在边数达数千万的图上可高效扩展至16个GPU。这些结果表明，基于SmartNIC的网络内预测是分区和压缩技术的实用补充，适用于大规模通信高效的全图GNN训练。
英文摘要
Full-graph GNN training delivers high accuracy but scales poorly on multi-server clusters due to heavy, irregular inter-node embedding exchanges. We present SNI-GNN, a SmartNIC-assisted full-graph training system that reduces communication while preserving accuracy by predicting remote embeddings in-network. SNI-GNN deploys a lightweight linear-trend predictor on SmartNICs to refine cached historical embeddings, coupled with an importance-based boundary-node sampling policy and an asynchronous DPU--GPU data pipeline with intermediate-result reuse. We provide error and convergence bounds showing that predictor bias remains controlled under bounded second-order dynamics and yields standard non-convex convergence with inexact gradients. Implemented on NVIDIA BlueField-3, SNI-GNN integrates with state-of-the-art full-graph systems, cuts communication by 21--45\%, achieves 1.3--3.6$\times$ end-to-end speedups over BNS-GCN and up to 1.29$\times$ over baseline SANCUS, with accuracy loss $\leq 0.01$, and scales efficiently to 16 GPUs on graphs with up to tens of millions of edges. These results indicate SmartNIC-based in-network prediction is a practical complement to partitioning and compression techniques for communication-efficient full-graph GNN training at scale.
