Architectural Implications of Embedding Dimension during GCN on CPU and GPU
This work addresses architectural optimization for GCN inference, which is incremental as it builds on existing hardware comparisons without introducing new methods.
The study analyzed the performance of Graph Convolutional Networks (GCNs) on CPUs and GPUs, focusing on how graph size, embedding dimension, and sampling affect inference, revealing that GPUs may not always be optimal due to sparsity and memory constraints.
Graph Neural Networks (GNNs) are a class of neural networks designed to extract information from the graphical structure of data. Graph Convolutional Networks (GCNs) are a widely used type of GNN for transductive graph learning problems which apply convolution to learn information from graphs. GCN is a challenging algorithm from an architecture perspective due to inherent sparsity, low data reuse, and massive memory capacity requirements. Traditional neural algorithms exploit the high compute capacity of GPUs to achieve high performance for both inference and training. The architectural decision to use a GPU for GCN inference is a question explored in this work. GCN on both CPU and GPU was characterized in order to better understand the implications of graph size, embedding dimension, and sampling on performance.