Sagitta: An Energy-Efficient Sparse 3D-CNN Accelerator for Real-Time 3D Understanding

• Published in: IEEE internet of things journal 2023-08, p.1-1
• Main Authors: Zhou, Changchun, Liu, Min, Qiu, Siyuan, Cao, Xugang, Fu, Yuzhe, He, Yifan, Jiao, Hailong
• Format: Article
• Language: English
• Subjects: Computer architecture; Convolution; edge computing; Energy efficiency; Internet of Things; Kernel; sparsity; speedup; Task analysis; three-dimensional convolutional neural networks; Three-dimensional displays; Three-dimensional inference
• ISSN: 2327-4662
• DOI: 10.1109/JIOT.2023.3306435

Three-dimensional (3D) understanding or inference has received increasing attention, where 3D convolutional neural networks (3D-CNNs) have demonstrated superior performance compared to two-dimensional CNNs (2D-CNNs), since 3D-CNNs learn features from all three dimensions. However, 3D-CNNs suffer from intensive computation and data movement. In this paper, Sagitta, an energy-efficient low-latency on-chip 3D-CNN accelerator, is proposed for edge devices. Locality and small differential value dropout are leveraged to increase the sparsity of activations. A full-zero-skipping convolutional microarchitecture is proposed to fully utilize the sparsity of weights and activations. A hierarchical load-balancing scheme is also introduced to increase the hardware utilization. Specialized architecture and computation flow are proposed to enhance the effectiveness of the proposed techniques. Fabricated in a 55-nm CMOS technology, Sagitta achieves 3.8 TOPS/W for C3D at a latency of 0.1 s and 4.5 TOPS/W for 3D U-Net at a latency of 0.9 s at 100 MHz and 0.91 V supply voltage. Compared to the state-of-the-art 3D-CNN and 2D-CNN accelerators, Sagitta enhances the energy efficiency by up to 379.6× and 11×, respectively.