15.2 A 28nm 64Kb Inference-Training Two-Way Transpose Multibit 6T SRAM Compute-in-Memory Macro for AI Edge Chips

• Main Authors: Su, Jian-Wei, Si, Xin, Chou, Yen-Chi, Chang, Ting-Wei, Huang, Wei-Hsing, Tu, Yung-Ning, Liu, Ruhui, Lu, Pei-Jung, Liu, Ta-Wei, Wang, Jing-Hong, Zhang, Zhixiao, Jiang, Hongwu, Huang, Shanshi, Lo, Chung-Chuan, Liu, Ren-Shuo, Hsieh, Chih-Cheng, Tang, Kea-Tiong, Sheu, Shyh-Shyuan, Li, Sih-Han, Lee, Heng-Yuan, Chang, Shih-Chieh, Yu, Shimeng, Chang, Meng-Fan
• Format: Conference Proceeding
• Language: English
• Subjects: Common Information Model (computing); Machine learning; SRAM cells; Threshold voltage; Training
• ISSN: 2376-8606
• DOI: 10.1109/ISSCC19947.2020.9062949

Many Al edge devices require local intelligence to achieve fast computing time (t AC ), high energy efficiency (EF), and privacy. The transfer-learning approach is a popular solution for Al edge chips, wherein data used to re-train the Al in the cloud is used to fine-tune (re-train) a few of the neural layers in edge devices. This enables the dynamic incorporation of data from in-situ environments or private information. Computing-in-memory (CIM) is a promising approach to improve EF for Al edge chips, existing CIM schemes support inference [1]-[5] with forward (FWD) propagation; however, they do not support training, requiring both FWD and backward (BWD) propagation, due to differences in weight-access flow for FWD and BWD propagation. As Fig. 15.2.1 shows, efforts to increase the precision of the input (IN), weight (W), and/or output (OUT) tend to degrade r AC and EF for training operations irrespective of scheme: digital FWD and BWD (DF-DB) or CIM-FWD-digital-BWD (CiMF-DB). This work develops a two-way transpose (TWT) SRAM-CIM macro supporting multibit MAC operations for FWD and BWD propagation with fast r AC and high EF within a compact area. The proposed scheme features (1) A TWT multiply cell (TWT-MC) with a high resistance to process variation; and (2) a small-offset gain-enhancement sense amplifier (SOGE-SA) to tolerate a small read margin. A 28nm 64Kb TWT SRAM-CIM macro was fabricated using a foundry-provided compact 6T-SRAM cell for SRAM-CIM devices supporting both inference and training operations for the first time. This macro also demonstrates the fastest t AC (3.8 - 21ns) and highest EF (7 - 61.1TOPS/w) for MAC operations using 2 - 8b inputs, 4 - 8b weights and 12 − 20b outputs.