Energy Minimization Computing using Field Coupled Computing with Nanomagnets

Energy Minimization Computing using Field Coupled Computing with Nanomagnets

Simulation and experimental validation of arrays of nanomagnet cells.

Various configurations of Multi-Layer Cells for clocking of nanomagnet cells.

Computational cost of traditional vision computing as a function of problem size (red plot) and the cost with proposed computing using nanomagnets (blue plot)

Comparison of quality of the nano-magnet solution with respect to traditional software based solutions

This work was supported in part by the National Science Foundation under grant CCF 0829838. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).

Sudeep Sarkar

Professor, Computer Science and Engineering

Sanjukta Bhanja

Associate Professor, Electrical Engineering

Javier Puleico

PhD, 2011 now at BSNL

Anita Kumari

PhD, 2011

Ravi Panchumarthy

PhD Candidate, 2011

J. Das, S. Alam, S. Bhanja, “Low Power CMOS-Magnetic Nano-Logic With Increased Bit Controllability,” IEEE Nanotechnology Conference, 2011.

J.F. Pulecio, S. Bhanja, and S. Sarkar, “Experimental Demonstration of Viability of Energy Minimizing Computing using Nano-magnets,” IEEE Nanotechnology Conference, 2011.

R. Panchumarthy, D. Karunaratne, S. Sarkar, S. Bhanja, “Tool for Analysis and Quantification of Fabrication Layouts in Nanomagnet-based Computing,” IEEE Nanotechnology Conference, 2011.

J. Das, S. M. Alam, and S. Bhanja, “Low Power Magnetic Quantum Cellular Automata Realization Using Magnetic Multi-Layer Structures,” IEEE Journal on Emerging and Selected Topics in Circuits and Systems (JETCAS), accepted for publication, June 2011.

J. F. Pulecio, P. Pendru, A. Kumari, S. Bhanja, "Magnetic Cellular Automata Wire Architectures," IEEE Transactions on Nanotechnology, vol. PP, no. 99, pp. 1, 0

A.Kumari, S. Bhanja, "Landauer Clocking for Magnetic Cellular Automata (MCA) Arrays," IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 19, no. 4, pp. 714-717, April 2011

A.Kumari, S. Sarkar, J. F. Pulecio, D. Karunaratne, S. Bhanja, "Study of magnetization state transition in closely spaced nanomagnet two-dimensional array for computation," Journal of Applied Physics , vol. 109, no. 7, pp. 07E513-07E513-3, Apr 2011,

J. F. Pulecio, S. Bhanja, "Magnetic cellular automata coplanar cross wire systems," Journal of Applied Physics , vol. 107, no. 3, pp. 034308-034308-5, Feb 2010

D. K. Karunaratne, J. F. Pulecio, S. Bhanja, "Driving magnetic cells for information storage and propagation," IEEE Nanotechnology Materials and Devices Conference (NMDC), pp. 360-363, 12-15 Oct. 2010

A.Kumari, S. Bhanja, "Magnetic Cellular Automata (MCA) arrays under spatially varying field," IEEE Nanotechnology Materials and Devices Conference, pp. 50-53, 2-5 June 2009

J. F. Pulecio, S. Bhanja, "Magnetic Cellular Automata wires," Nanotechnology Materials and Devices Conference, 2009. NMDC '09. IEEE , vol., no., pp.73-75, 2-5 June 2009

A. Kumari, J. F. Pulecio, S. Bhanja, S.; , "Defect characterization in magnetic field coupled arrays," Quality of Electronic Design, 2009. ISQED 2009. Quality Electronic Design , vol., no., pp.436-441, 16-18 March 2009

S. Sarkar, S. Bhanja, “Direct Quadratic Minimization Using Magnetic Field-Based Computing,” pp.31-34, IEEE International Workshop on Design and Test of Nano Devices, Circuits and Systems, 2008

S. Sarkar and S. Bhanja, “Synthesizing energy minimizing quantum-dot cellular automata circuits for vision computing,” IEEE Conference on Nanotechnology, vol. 2, pp. 541-544, 2005.

Goal

Background

Driven by needs for very dense storage (HDD) and memory devices (MRAM), there have been tremendous advances in fabrication methods for patterning magnetic media at nanoscales. It is now possible to create magnets below 100nm dimensions. Such magnets can be fairly well approximated by single magnetic domain models. By exploiting geometric and material anisotropy, one can construct nanomagnet that have two stable magnetic arrangements, which can be used to represent 0 and 1, or in the absence of asymmetry a continuous state representation. The size of the nanomagnets in such devices are limited not by technology, but by the need to isolate the devices from each other. This dipolar interaction between neighboring magnetic devices can be exploited for computing. So far, there have been suggestions for using them in Boolean logic based computing.

We are investigating how to harness the energy minimizing aspect of magnetic operations to directly solve quadratic optimization problems. We call this Magnetic Field-based Computing (MFC); it uses an interacting system of single-domain nanomagnets. Since nano-devices will be associated with high error rates, both at fabrication and during operations, it makes sense to consider error-tolerant applications, where the cost of failure of not finding the optimal solution is not high; even solutions that are close to optimal ones suffices in practice. One such context is in quadratic optimization that arises in object recognition problems in computer vision. Unlike logic and arithmetic computing tasks that demand exact computations, vision problems can work with near optimal solutions. These vision problems place high demand on computational resources (on Boolean logic based computing platforms).