AnalogicLab Site v0.1

Add to favorites

Make us your homepage


	Mission

	News

	Contact Us

	Memberlist

	Downloads




	Publications (9)

	Projects & Highlights (8)

	Research Grants (8)

	Educational Activities (8)

	Memberships (8)




	CNN tutorial

	Partners

	CNN Bibliography

In the era of multi billion transistor chips, thousands of million transistor processors can be placed on a single chip. Indeed, they are already in the market (1024-DSP FPGAs, 1024-core Graphics processors, as well as 25 kilo-core visual microprocessors, to name just a few). Since the beginning of the new century, a new kind of Moore’s law is emerging, the doubling of the number of processor cores in a single chipWhat is dramatically missing is not only an operating syystem or a language to handle the mega-processor computers made of these kilo-core chips, however, indeed, the algorithmic thinking is missing. In this, really different, computer architectures each individual processor core has a geometric address. Namely, due to the physical constraints of the nano-scale CMOS technology (wire delay is increasingly bigger than gate delay), the communication speed between the cores (and memory banks) depends heavily on the physical distance between the cores. This is much more than parallelism, this is a cellular constraint where the local, neighbor interaction has a speed preference over the global, farther communication speed. Hence, the geometrical configuration of the processor cores plays an important role. In this sense these are cellular morphic computers.

In this new direction, our Cellular and Sensory Wave Computer Research Laboratory has been working since more than 15 years, pioneering „CNN Technology”, the core discipline and earlier prototypes of visual microprocessor Chips hosting many thousand processors and optical sensors on each processor. The essence of the discipline and technology can be found in two recent publications and a textbook:

T. Roska, “Circuits, Computers, and beyond Boolean Logic”, Int. J. Circuit Theory and Appl., vol. 35, No. 3.

T. Roska, “Cellular Wave Computers for Nano-Tera-scale technology – beyond Boolean, spatial-temporal logic in million processor devices”, Electronics Letters, vol.43,, pp. 427-429, 2007 (Insight Letter)

L.O.Chua and T. Roska, Cellular Neural Networks and Visual Computing, Cambridge University Press, Cambridge, UK, 2002, 2005

and the company web sites: www.anafocus.com and www.eutecus.com

Hence, the research activity in Laboratory is organized in relation with the original paradigm of the Cellular Sensory Wave Computer, in close collaboration since many years with key laboratories at the University of California, Berkeley (Leon O. Chua) and the University of Seville (Ángel Rodríguez-Vázquez), and other groups worldwide. The main subjects are as follows:

Qualitative and realizability theory of Cellular Wave Computers
Portable integrated sensory computers using chips of thousands of processors
Cellular 3D super-computers
Algorithms and platforms for multi-modal, real-time fusion and navigation
Programmable optical cellular computers

A strong link has been established with Neuromorphic and Bio-inspired models, especially related to the multi-channel retinal models and adaptivity /plasticity.

Historically, it is interesting that the two pioneers of computing, Alan Turing and John von Neumann both made prophetical suggestions in this direction. The Morphogenesis paper of Turing and the Cellular Automaton paper of Von Neumann, made the first steps in this direction.

MTA - access
to CELL