Defect characterization in magnetic field coupled arrays Kumari, A. Pulecio, J.F. Bhanja, S. |
 |
|
|
| Abstract Magnetic Cellular Automata (MCA) utilizes mutual exchange energies of neighboring magnetic cells to order the single-domain magnetic cell which in turn performs computational tasks. In this paper, we study three dominant type of geometric defects (missing, spacing, merging) in array (used as interconnects) based on our fabrication experiments. We study effect of these defects in three segments of the array (near-input, center and near-output) and we have observed that location of these defects play an important role in masking of the errors. The observed simulation results indicate that most of the defects occurring around center and near-output would be masked generating correct behavior while defects in the near-input segment would mostly cause erroneous output. We also observe that MCA is extremely robust towards space irregularities, one of the most common form of defect we observed through our fabrication techniques. |
Monday, March 16, 2009
Defect characterization in magnetic field coupled arrays
Friday, January 16, 2009
An Error Model to Study the Behavior of Transient Errors in Sequential Circuits
An Error Model to Study the Behavior of Transient Errors in Sequential Circuits Lingasubramanian, K. Bhanja, S. |
|
|
|
| Abstract In sequential logic circuits the transient errors that occur in a particular time frame will propagate to consecutive time frames thereby making the device more vulnerable. In this work we propose a probabilistic error model for sequential logic that can measure the expected output error probability, given a probabilistic input space, that account for both spatial dependencies and temporal correlations across the logic, using a time evolving causal network. We demonstrate our error model using MCNC and ISCAS benchmark circuits and validate it with HSpice simulations. Our observations show that, significantly low individual gate error probabilities produce at least 5 fold higher output error probabilities. The average error percentage of our results with reference to HSpice simulation results is only 4.43%. Our observations show that the order of temporal dependency of error varies for different sequential circuits. |
Tuesday, September 16, 2008
Direct Quadratic Minimization Using Magnetic Field-Based Computing
Direct Quadratic Minimization Using Magnetic Field-Based Computing Sarkar, S. Bhanja, S. |
|
|
|
| Abstract We explore an unconventional front in computing,which we call magnetic field-based computing (MFC), that harnesses the energy minimization aspects of a collection of nanomagnets to solve directly quadratic energy minimization problems, such as those arising in computationaolly intensive computer vision tasks. The Hamiltonian of a collection of bipolar nanomagnets is governed by the pairwise dipolar interactions.The ground state of a nanomagnet collection minimizes this Hamiltonian. We have devised a computational method, based on multi-dimensional scaling, to decide upon the spatial arrangement of nanomagnets that matches a particular quadratic minimization problem. Each variable is represented by a nanomagnet and the distances between them are such that the dipolar interactions match the corresponding pairwise energy term in the original optimization problem. We select the nanomagnets that participate in a specific computation from a field of regularly placed nanomagnets. The nanomagnets that do not participate are deselected using transverse magnetic fields. We demonstrate these ideas by solving Landau-Lifshitz equations as implemented in the NISTpsilas micro-magnetic OOMMF software. |
|
|
|
Index Terms |