9.1 A 7nm 4-Core AI Chip with 25.6TFLOPS Hybrid FP8 Training, 102.4TOPS INT4 Inference and Workload-Aware Throttling

• Published in: 2021 IEEE International Solid- State Circuits Conference (ISSCC) Vol. 64; pp. 144 - 146
• Main Authors: Agrawal, Ankur, Lee, Sae Kyu, Silberman, Joel, Ziegler, Matthew, Kang, Mingu, Venkataramani, Swagath, Cao, Nianzheng, Fleischer, Bruce, Guillorn, Michael, Cohen, Matthew, Mueller, Silvia, Oh, Jinwook, Lutz, Martin, Jung, Jinwook, Koswatta, Siyu, Zhou, Ching, Zalani, Vidhi, Bonanno, James, Casatuta, Robert, Chen, Chia-Yu, Choi, Jungwook, Haynie, Howard, Herbert, Alyssa, Jain, Radhika, Kar, Monodeep, Kim, Kyu-Hyoun, Li, Yulong, Ren, Zhibin, Rider, Scot, Schaal, Marcel, Schelm, Kerstin, Scheuermann, Michael, Sun, Xiao, Tran, Hung, Wang, Naigang, Wang, Wei, Zhang, Xin, Shah, Vinay, Curran, Brian, Srinivasan, Vijayalakshmi, Lu, Pong-Fei, Shukla, Sunil, Chang, Leland, Gopalakrishnan, Kailash
• Format: Conference Proceeding
• Language: English
• Published: IEEE 13-02-2021
• Subjects: AI accelerators; Inference algorithms; Integrated circuit modeling; Optimization; Power system management; Solid state circuits; Training
• ISSN: 2376-8606
• DOI: 10.1109/ISSCC42613.2021.9365791

Low-precision computation is the key enabling factor to achieve high compute densities (TOPS/W and TOPS/mm 2 ) in AI hardware accelerators across cloud and edge platforms. However, robust deep learning (DL) model accuracy equivalent to high-precision computation must be maintained. Improvements in bandwidth, architecture, and power management are also required to harness the benefit of reduced precision by feeding and supporting more parallel engines to achieve high sustained utilization and optimize performance within a given product power envelope. In this work, we present a 4-core AI chip in 7nm EUV technology that exploits cutting-edge algorithmic advances for iso-accurate models in low-precision training and inference [1, 2] and aggressive circuit/architecture optimization to achieve leading-edge power-performance. The chip supports fp16 (DLFIoat16 [8]) and hybrid-fp8 (hfp8) [1] formats for training and inference of DL models, as well as int4 and int2 formats for highly scaled inference.