An Isotropic Shift-Pointwise Network for Crossbar-Efficient Neural Network Design

Ziyi Guan1; Boyu Li1; Yuan Ren1; Muqun Niu1; Hantao Huang2; Graziano Chesi1; Hao Yu2; Ngai Wong1

2024-03-27

1530-1591

979-8-3503-4860-6

2024 Design, Automation & Test in Europe Conference & Exhibition (DATE)

25-27 March 2024

Valencia, Spain

Resistive random-access memory (RRAM), with its programmable and nonvolatile conductance, permits compute-in-memory (CIM) at a much higher energy efficiency than the traditional von Neumann architecture, making it a promising candidate for edge AI. Nonetheless, the fixed-size crossbar tiles on RRAM are inherently unfit for conventional pyramid-shape convolutional neural networks (CNNs) that incur low crossbar utilization. To this end, we recognize the mixed-signal (digital-analog) nature in RRAM circuits and customize an isotropic shift-pointwise network that exploits digital shift operations for efficient spatial mixing and analog pointwise operations for channel mixing. To fast ablate various shift-pointwise topologies, a new recon-figurable energy-efficient shift module is designed and packaged into a seamless mixed-domain simulator. The optimized design achieves a near-100% crossbar utilization, providing a state-of-the-art INT8 accuracy of 94.88% (76.55%) on the CIFAR-10 (CIFAR-100) dataset with 1.6M parameters, which sets a new standard for RRAM-based AI accelerators.

其他

1.Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong
2.School of Microelectronics, Southern University of Science and Technology, Shenzhen

Ziyi Guan,Boyu Li,Yuan Ren,et al. An Isotropic Shift-Pointwise Network for Crossbar-Efficient Neural Network Design[C],2024.