News Archives

RSS Feed

  • UNM
  • >Home
  • >News
  • >2006
  • >March
  • >Towards Novel Computing Paradigms: Computation with Irregular and Imperfect Cellular Assemblies

Towards Novel Computing Paradigms: Computation with Irregular and Imperfect Cellular Assemblies

March 28, 2006

  • Date: Tuesday, March 28, 2006 
  • Time: 11:00 am — 12:15 pm
  • Place: Woodward 149

Christof Teuscher
(UNM Faculty Candidate) Los Alamos National Laboratory

Novel materials and fabrication technologies bear unique opportunities for self-assembling tomorrow’s multi-billion component computing machines in a largely random manner, which would lower fabrication costs significantly compared to today’s definite ad hoc assemblies. However, we do not possess the formalisms, methodologies, and tools do deal with such unconventional machines. This challenge—at the interface between biology, computer science, material science, and physics—needs to be addressed in a very interdisciplinary way, and, if successful, would likely allow for a quantum leap in performance of computing machines.

In this talk, I will first present an overview on some biologically inspired computing architectures, both with a regular and an irregular basic structure. Using cellular automata and random boolean networks as a showcase, I will illustrate that irregular assemblies have major advantages over regular ones for solving certain “global” tasks. I will further describe the properties of a physically plausible small-world interconnect fabric that is inspired by modern network on chip paradigms. The framework’s key parameters, such as the connectivity, the number of switch blocks, the number of virtual channels, the routing strategy, and the distribution of long- and short-range connections shall be varied while measuring the network’s transport characteristics and robustness against failures. The results show that irregular interconnect fabrics have major advantages in terms of performance and robustness over purely regular fabrics.

The last part of the talk shall be dedicated to future work and challenges in realizing higher-level adaptive and robust systems on top of novel computing substrates.

BIOGRAPHY: Christof Teuscher obtained his M.Sc. and Ph.D. degree in computer science from the Swiss Federal Institute of Technology in Lausanne (EPFL) in 2000 and 2004 respectively. He was a postdoctoral researcher at the University of California in San Diego (UCSD) from 2004 to 2005 and is currently a Director’s Postdoctoral Fellow at the Los Alamos National Laboratory. His main research interests include biologically-inspired computing, novel and unconventional computing architectures and paradigms, and complex & adaptive systems. Teuscher has received several awards and fellowships. For more information visit http://www.teuscher.ch/christof.