PruneInfer: Exact Full-Neighborhood GNN Inference of Large Datasets on a Single GPU via Full Topology Pruning
PruneInfer is the only system on the authors' experimental configuration that maintains accuracy comparable to full-graph inference across all evaluated models and achieves up to 4.49 × speedup over DGL and up to 7.05 × speedup over GDL-GNN across the evaluated datasets.
Li-Yang Wu, Meng-Han Jia, Chun-Ye Gong et al.
· Proceedings of the Internati... · 0 citations