Abstract:Emerging applications are extremely demanding in terms of storage, computing power and energy efficiency. On the other hand, both today’s computer architectures and device technologies are facing major challenges making them incapable to deliver the required functionalities and features. All of these have led to the slowdown of the traditional device scaling. In order for computing systems to continue deliver sustainable benefits for the foreseeable future society, alternative computing architectures and notions have to be explored in the light of emerging new device technologies.
This talk will address the potential of using emerging memrisive devices for computation-in-memory. First, the challenges that today’s architecture and technology are facing will be discussed; this will motivate not only the need of new architectures, but also the need of new technologies enabling the implementation of such architecture in efficient manner. Next, the concept of computation-in-memory (CIM) will be defined and the differences with other computing paradigms will be highlighted. Then, a classification of CIM architectures will be provided, as well as their impact of the circuit design. Thereafter, using the basic fundamentals of memristive devices and their potential for memory, logic and arithmetic designs, and how they enable CIM architecture will be explored. In addition, the huge potential of such architecture in realizing order of magnitude improvement in terms of computing and energy efficiency is illustrated by comparing it with the state-of-the art architectures for different applications. Finally the research directions in computation-in-memory are highlighted.
Said Hamdioui (http://www.ce.ewi.tudelft.nl/hamdioui/) is currently Chair Professor on Dependable and Emerging Computer Technologies, Head of the Computer Engineering Laboratory (CE-Lab), and Head of the department of Quantum and Computer Engineering of the Delft University of Technology, the Netherlands. He received the MSEE and PhD degrees (both with honors) from TUDelft. Prior to joining TUDelft as a professor, Hamdioui spent over seven years within industry including at Intel (Califorina, USA), Philips Semiconductors R&D (Crolles, France) and Philips/ NXP Semiconductors (Nijmegen, The Netherlands). His research focuses on two domains: Dependable CMOS nano-computing (including Reliability, Testability, Hardware Security) and emerging technologies and computing paradigms (including 3D stacked ICs, memristors for logic and storage, in-memory-computing).
Hamdioui owns two patents, has published one book and contributed to other two, and had co-authored over 180 conference and journal papers. Hamdioui is also the initiator and the coordinator of an EU project on computation-in-memory architecture based on resistive devices (http://www.mnemosene.eu/). He is co-founder of CognitiveIC (a startup on hardware dependability). He is the recipient of many international/national awards, has consulted for many semiconductor companies, and has served on the editorial board of different journals.
downloadFile