Category Archives: Electronic Systems Research

About This Lab

Adapt or perish, now as ever, is nature’s inexorable imperative – H. G. Wells

With frontier silicon technology scaling, current and future generations of electronic computing systems are expected to be highly parallel with many interconnected heterogeneous cores. The emergence of such systems is likely to enable performance and scalability at unprecedented levels, and provide with the opportunities for many new and concurrent applications. However, increased device-level power densities in these system will render prohibitive energy consumption with extremely large operating costs and operating temperatures. Hence, ensuring energy efficiency and reliability at low-cost, while also enabling new and concurrent applications is highly challenging.

In Adaptive lab, we investigate into modeling, prototyping and implementation of low-cost design and runtime methodologies to tackle the above energy efficiency and reliability challenges. The overall aim is to learn how each application exercises the underlying hardware architectural components and thus exploit every opportunity to: 1) prune the unnecessary energy consumption, and 2) control the power knobs of the useful architectural components in an optimised way to ensure significant improvement in the scalability, energy-efficiency and reliability for these systems.

A wordmap of the keywords relevant to the various aspects of our research can be found below:

wordle 2

Specifically, our current research interests are:

A. Energy-efficient and Reliable Electronic Systems Design: using hardware/software co-design and co-optimisation; programming models (OpenMP, OpenCL, OpenCV) and middleware/firmware designs.

B. Approximate Computing Systems Design: using variable exponent arithmetic and adaptive programming models for large scale systems, such as big data applications, and cloud server systems.

C. Software/Hardware Co-design for Emerging/Translational Applications: Custom hardware and firmware design of translational applications and deep-learning systems with the aim of achieving required energy efficiency and performance.

D. Energy-efficient High-performance Computing: Hardware systems, architectures and programming models of energy-efficient high-performance computing systems.

If you are interested in the research and development of the above or similar areas, and would like to get involved as a new postgraduate (i.e. PhD) researcher or intern, please contact Dr. Rishad Shafik (e-mail: rishad <dot> shafik <at> Newcastle <dot> ac <dot> uk) with your CV and research statement.

More details regarding Dr. Rishad Shafik can be found here.