*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Atlanta, GA | Posted: May 10, 2021
School of Electrical and Computer Engineering
Summary Sentence:
Jong-Hyeok Yoon, Muya Chang, and Arijit Raychowdhury won the best regular paper award at the IEEE Custom Integrated Circuits Conference (CICC) 2021.
Full Summary:
Jong-Hyeok Yoon, Muya Chang, and Arijit Raychowdhury won the best regular paper award at the IEEE Custom Integrated Circuits Conference (CICC) 2021.
Jong-Hyeok Yoon, Muya Chang, Arijit RaychowdhuryJong-Hyeok Yoon, Muya Chang, and Arijit Raychowdhury won the best regular paper award at the IEEE Custom Integrated Circuits Conference (CICC) 2021. The conference was held April 25-30 in a virtual format.
The title of the award-winning paper is “A 40nm 100Kb 118.44TOPS/W Ternary-weight Compute-in-Memory RRAM Macro with Voltage-sensing Read and Write Verification for Reliable Multi-bit RRAM Operation.” The explosive growth of memory-centric workloads, from artificial intelligence to graph processing to data-analytics, has necessitated research and development of next-generation memory sub-systems. In particular, resistive RAM (RRAM) has emerged as a promising candidate because of its high density, high read speed, and non-volatility.
This paper presented a 101.4Kb bit-bit encoding in an RRAM macro on a 40nm CMOS node, supporting compute-in-memory for ternary weight networks. A key innovation is the design and verification of a novel voltage-based read with active feedback that addresses a key technology challenge – a low resistance ratio between the high resistance state and the low resistance state in high-endurance RRAM. The memory macro also features iterative write with verification to facilitate a reliable multi-bit encoding under a sense margin narrow margin. This is one of the first demonstrations of reliable multi-bit encoding on RRAM in a foundry process on a scaled technology node. The fabrication was supported by the Taiwan Semiconductor Manufacturing Company’s (TSMC) university shuttle services.
Raychowdhury leads the Integrated Circuits and Systems Research Lab (ICSRL), where this research was conducted in collaboration with TSMC. He holds the Motorola Solutions Foundation Professorship in the Georgia Tech School of Electrical and Computer Engineering (ECE). Yoon served as a postdoctoral researcher in ICSRL, and recently began work as an assistant professor at Daegu Gyeongbuk Institute of Science and Technology (DGIST) in the Republic of Korea. Chang recently graduated with his Ph.D. under the direction of Dr. Raychowdhury and is now a postdoctoral researcher in ICSRL. Vince Khwa, Y.D. Chen, and Marvin Chang are co-authors on the paper and are all employees of TSMC.
