ReRAM-based graph attention network with node-centric edge searching and hamming similarity
- Resource Type
- Conference
- Authors
- Mao, Ruibin; Sheng, Xia; Graves, Catherine; Xu, Cong; Li, Can
- Source
- 2023 60th ACM/IEEE Design Automation Conference (DAC) Design Automation Conference (DAC), 2023 60th ACM/IEEE. :1-6 Jul, 2023
- Subject
- Components, Circuits, Devices and Systems
Computing and Processing
Engineering Profession
Design automation
Simulation
Graphics processing units
Hardware
Robustness
Energy efficiency
graph
graph attention network
sparse
ReRAM
memristor
crossbar array
- Language
The graph attention network (GAT) has demonstrated its advantages via local attention mechanism but suffered from low energy and latency efficiency when implemented on conventional von-Neumann hardware. This work proposes and experimentally demonstrates an algorithm-hardware co-designed GAT that runs efficiently and reliably in ReRAM-based hardware. The neighborhood information is retrieved from trained node embeddings stored on crossbars in a single time step, and attention is implemented by efficient hashing and hamming similarity for higher robustness. Our scaled simulation based on the experimentally-validated model shows only 0.9% accuracy loss with over 35,500x energy improvement on the Cora dataset compared with GPU, and 1.1% accuracy improvement with 2× energy improvement compared with state-of-the-art ReRAM-based GNN accelerator.