Thesis defence H.A. Du Nguyen: memristive devices

13 September 2019 10:00 - Location: Aula, TU Delft - By: webredactie

Computation-in-Memory based on Memristive Devices. Promotor: S. Hamdioui (EWI).

In-memory computing is a promising computing paradigm due to its capability to alleviate the memory bottleneck. It has even higher potential when implemented using memristive devices or memristors with various beneficial characteristics such as nonvolatility, high scalability, near-zero standby power consumption, high density, and CMOS compatibility. Exploring in-memory computing architectures in the combination with memristor technology is still in its infancy phase. Therefore, it faces challenges with respect to the development of the devices, circuits, architectures, compilers and applications. This thesis focuses on exploring and developing in-memory computing in terms of architectures (including classification, limited schemes of instruction set, micro-architecture, communication and controller, as well as automation and simulator), and circuits (including logic synthesis flow and interconnect network schemes). The work contains three main parts: In-Memory architecture classification and survey, two architectures representing two classes from the above classification, logic synthesis automation tools and interconnect networks used in these architectures. The architecture as well as simulation platform are in their infancy stage, and more efforts are required to fully utilize the architecture's potentials on big data applications. However, the analytical and simulation results show that the proposed architectures outperform the conventional architecture at least one order of magnitude in terms of performance and energy.

More information?

For access to theses by the PhD students you can have a look in TU Delft Repository, the digital storage of publications of TU Delft. Theses will be available within a few weeks after the actual thesis defence.