[2020-02-26] Dr. Shuo-Han Chen(陳碩漢), Academia Sinica, "One-Memory Computer System: When the Line Between Main Memory and Storage is Blurred via NVRAM"

Poster:Post date:2020-02-20
Title: One-Memory Computer System: When the Line Between Main Memory and Storage is Blurred via NVRAM
Date: 2020-02-26 11:00am-12:00pm
Location: R210, CSIE
Speaker: Dr. Shuo-Han Chen(陳碩漢), Academia Sinica
Hosted by: Prof. Chi-Sheng (Daniel) Shih


The growing computational power of today's computer systems demands larger DRAM-based main memory and faster NAND flash-based storage. However, both of today's DRAM and NAND flash technologies face their limitations in delivering high-density main memory and high-performance storage devices. The limitations of DRAM include the density limitation and the high idle energy consumption issue, while the NAND flash technology faces the difficulties in delivering high reliability, high speed, and high durability at the same time. For those energy and space-constraint computer systems (i.e., mobile devices), the limitations of DRAM and NAND flash are further exaggerated as the sizes of both the main memory and storage have grown multiple times over the years. To do away the limitations of DRAM and NAND flash, nonvolatile random access memory (NVRAM) is considered as a promising solution with its nice features, which are byte-addressability, nonvolatility, and near-zero idle energy consumption. Although directly deploying NVRAM to replace DRAM and NAND flash is workable, our investigation shows that this approach induces redundancy and cannot fully bring the benefits of NVRAM into computer systems. To resolve this issue, we proposed three new design considerations and critical changes for NVRAM-based computer systems, in which both the main memory and storage are hosted on NVRAM. This talk will cover the following aspects, including the memory-related issues faced by today's computer systems, the characteristics of NVRAM, our stretch goal for NVRAM, and the proposed one-memory computer system. Please feel free to join, and any forms of interaction are welcome during the talk.

Dr. Shuo-Han Chen is currently a postdoctoral fellow at the Institute of Information Science, Academia Sinica, Taiwan, since Mar. 2018. During his postdoctoral fellowship, he was invited to become a visiting scholar at the Department of Electrical and Computer Engineering, Purdue University, USA (2019/07 - 2019/12) and the Department of Computer Science and Engineering, the Chinese University of Hong Kong (CHUK), Hong Kong (2018/04 - 2018/12). Dr. Chen received his doctoral degree from the Department of Computer Science, National Tsing Hua University, Taiwan (2015/09 - 2018/03). Before his doctoral degree, his related experiences in IT industries include internships at ASUSTeK Computer, Broadcom Cooperation, and Telematicus at the United Kingdom. Dr. Chen's primary research interests include emerging non-volatile memory and storage technologies, memory and storage systems, next-generation memory/storage architecture designs, and low-power system designs for artificial neural networks.

Last modification time:2020-02-20 AM 10:33

cron web_use_log