G-MAP: A Graph-Neural Network Based Framework for Memory Access Prediction
Published in IEEE HPEC, 2023
In this paper, we introduce G-MAP, a novel Graph Neural Network-based framework for Memory Access Prediction. First, we propose Mem2Graph, a novel approach mapping a memory access sequence to a graph representation, capturing both the spatial and temporal locality in the memory access sequence which most existing methods fail to do. Second, we implement various GNNs for G-MAP, including Graph Convolutional Network (GCN), Gated Graph Sequence Neural Network (GG-NN), and Graph Attention Network (GAT).
Recommended citation: A. R. Gorle, P. Zhang, R. Kannan and V. K. Prasanna, "G-MAP: A Graph Neural Network-Based Framework for Memory Access Prediction," 2023 IEEE High Performance Extreme Computing Conference (HPEC), Boston, MA, USA, 2023, pp. 1-7, doi: 10.1109/HPEC58863.2023.10363605. https://ieeexplore.ieee.org/document/10363605