Accelerating Sparse Graph Neural Network Computation via Dense Tensor Core on GPUs
, Ph.D. candidate, University of California, Santa Barbara
Recently, graph neural networks (GNNs), as the backbone of graph-based machine learning, have shown great success in various domains (e.g., ecommerce). However, GNN performance is usually unsatisfactory due to the highly sparse and irregular graph-based operations. We'll give an in-depth analysis of the sparse operations in mainstream GNN, and explore the potential of using the dense Tensor Core in sparse graph neural network computation.