DGI: An Easy and Efficient Framework for GNN Model Evaluation
Open Access
- 4 August 2023
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 5439-5450
Abstract
While many systems have been developed to train graph neural networks (GNNs), efficient model evaluation, which computes node embedding according to a given model, remains to be addressed. For instance, using the widely adopted node-wise approach, model evaluation can account for over 90% of the time in the end-to-end training process due to neighbor explosion, which means that a node accesses its multi-hop neighbors. The layer-wise approach avoids neighbor explosion by conducting computation layer by layer in GNN models. However, layer-wise model evaluation takes considerable implementation efforts because users need to manually decompose the GNN model into layers, and different implementations are required for GNN models with different structures. In this paper, we present DGI -a framework for easy and efficient GNN model evaluation, which automatically translates the training code of a GNN model for layer-wise evaluation to minimize user effort. DGI is general for different GNN models and evaluation requests (e.g., computing embedding for all or some of the nodes), and supports out-of-core execution on large graphs that cannot fit in CPU memory. Under the hood, DGI traces the computation graph of GNN model, partitions the computation graph into layers that are suitable for layer-wise evaluation according to tailored rules, and executes each layer efficiently by reordering the computation tasks and managing device memory consumption. Experiment results show that DGI matches hand-written implementations of layer-wise evaluation in efficiency and consistently outperforms node-wise evaluation across different datasets and hardware settings, and the speedup can be over 1,000x.Keywords
Funding Information
- Guangdong Basic and Applied Basic Research Foundation (2021A1515110067)
- CUHK (4055146)
- Shenzhen Municipal Fundamental Research Program (20220815112848002)
This publication has 24 references indexed in Scilit:
- Heterogeneous Graph Attention Network2019
- Heterogeneous Graph Neural Networks for Malicious Account Detection2018
- Graph Convolutional Neural Networks for Web-Scale Recommender Systems2018
- Knowledge Transfer for Out-of-Knowledge-Base Entities : A Graph Neural Network Approach2017
- Speedup Graph Processing by Graph Ordering2016
- Rabbit Order: Just-in-Time Parallel Reordering for Fast Graph Analysis2016
- Defining and evaluating network communities based on ground-truthKnowledge and Information Systems, 2013
- Think Locally, Act Globally: Highly Balanced Graph Partitioning2013
- Learning spectral graph transformations for link prediction2009
- A linear time implementation of the reverse Cuthill-McKee algorithmBIT Numerical Mathematics, 1980