arXiv subjects

Kerem Yunus Camsari

Publications and source records attributed to Kerem Yunus Camsari.

7 recordsLinked to original sources

arXiv2018

Low-Energy Deep Belief Networks using Intrinsic Sigmoidal Spintronic-based Probabilistic Neurons

A low-energy hardware implementation of deep belief network (DBN) architecture is developed using near-zero energy barrier probabilistic spin logic devices (p-bits), which are modeled to realize an intrinsic sigmoidal activation function. A CMOS/spin based weighted array structure is designed to implement a restricted Boltzmann machine (RBM). Device-level simulations based on precise physics relations are used to validate the sigmoidal relation between the output probability of a p-bit and its input currents. Characteristics of the resistive networks and p-bits are modeled in SPICE to perform a circuit-level simulation investigating the performance, area, and power consumption tradeoffs of the weighted array. In the application-level simulation, a DBN is implemented in MATLAB for digit recognition using the extracted device and circuit behavioral models. The MNIST data set is used to assess the accuracy of the DBN using 5,000 training images for five distinct network topologies. The results indicate that a baseline error rate of 36.8% for a 784x10 DBN trained by 100 samples can be reduced to only 3.7% using a 784x800x800x10 DBN trained by 5,000 input samples. Finally, Power dissipation and accuracy tradeoffs for probabilistic computing mechanisms using resistive devices are identified.

cs.ET↗

arXiv2017

Hardware emulation of stochastic p-bits for invertible logic

The common feature of nearly all logic and memory devices is that they make use of stable units to represent 0's and 1's. A completely different paradigm is based on three-terminal stochastic units which could be called "p-bits", where the output is a random telegraphic signal continuously fluctuating between 0 and 1 with a tunable mean. p-bits can be interconnected to receive weighted contributions from others in a network, and these weighted contributions can be chosen to not only solve problems of optimization and inference but also to implement precise Boolean functions in an inverted mode. This inverted operation of Boolean gates is particularly striking: They provide inputs consistent to a given output along with unique outputs to a given set of inputs. The existing demonstrations of accurate invertible logic are intriguing, but will these striking properties observed in computer simulations carry over to hardware implementations? This paper uses individual micro controllers to emulate p-bits, and we present results for a 4-bit ripple carry adder with 48 p-bits and a 4-bit multiplier with 46 p-bits working in inverted mode as a factorizer. Our results constitute a first step towards implementing p-bits with nano devices, like stochastic Magnetic Tunnel Junctions.

cs.ET↗

arXiv2017

Stochastic p-bits for Invertible Logic

Conventional logic and memory devices are built out of deterministic units such as transistors, or magnets with energy barriers in excess of 40-60 kT. We show that stochastic units, p-bits, can be interconnected to create robust correlations that implement Boolean functions with impressive accuracy, comparable to standard circuits. Also they are invertible, a unique property that is absent in digital circuits. When operated in the direct mode, the input is clamped, and the network provides the correct output. In the inverted mode, the output is clamped, and the network fluctuates among possible inputs consistent with that output. We present an implementation of an invertible gate to bring out the key role of a three-terminal building block to enable the construction of correlated p-bit networks. The results for this implementation agree well with those from a universal model, showing that p-bits need not be magnet-based: any three-terminal tunable random bit generator should be suitable. We present an algorithm for designing a Boltzmann machine (BM) with symmetric connections that implements a given truth table. We then show how BM Full Adders can be interconnected in a partially directed manner to implement large operations such as 32-bit addition. Hundreds of p-bits get precisely correlated such that the correct answer out of 2^33 possibilities can be extracted by looking at the mode of a number of time samples. With perfect directivity a small number of samples is enough, while for less directed connections more samples are needed, but even in the former case invertibility is largely preserved. This combination of accuracy and invertibility is enabled by the hybrid design that uses bidirectional units to construct circuits with partially directed connections. We establish this result with examples including a 4-bit multiplier which in inverted mode functions as a factorizer.

cond-mat.mes-hall↗

arXiv2017

Low Barrier Nanomagnets as p-bits for Spin Logic

It has recently been shown that a suitably interconnected network of tunable telegraphic noise generators or "p-bits" can be used to perform even precise arithmetic functions like a 32-bit adder. In this paper we use simulations based on the stochastic Landau-Lifshitz-Gilbert (sLLG) equation to demonstrate that similar impressive functions can be performed using unstable nanomagnets with energy barriers as low as a fraction of a kT. This is surprising since the magnetization of low barrier nanomagnets is not telegraphic with discrete values of +1 and -1. Rather it fluctuates randomly among all values between -1 and +1, and the output magnets are read with a thresholding device that translates all positive values to 1 and all negative values to zero. We present sLLG-based simulations demonstrating the operation of a 32-bit adder with a network of several hundred nanomagnets, exhibiting a remarkably precise correlation: The input magnets {A} and {B} as well as the output magnets {S} all fluctuate randomly and yet the quantity A+B-S is sharply peaked around zero! If we fix {A} and {B}, the sum magnets {S} rapidly converge to a unique state with S=A+B so that the system acts as an adder. But unlike standard adders, the operation is invertible. If we fix {S} and {B}, the remaining magnets {A} converge to the difference A=S-B. These examples suggest a new direction for the field of nanomagnetics away from stable high barrier magnets towards stochastic low barrier magnets which not only operate with lower currents, but are also more promising for continued downscaling.

cond-mat.mes-hall↗

arXiv2017

Intrinsic optimization using stochastic nanomagnets

This paper draws attention to a hardware system which can be engineered so that its intrinsic physics is described by the generalized Ising model and can encode the solution to many important NP-hard problems as its ground state. The basic constituents are stochastic nanomagnets which switch randomly between the $\pm 1$ Ising states and can be monitored continuously with standard electronics. Their mutual interactions can be short or long range, and their strengths can be reconfigured as needed to solve specific problems and to anneal the system at room temperature. The natural laws of statistical mechanics guide the network of stochastic nanomagnets at GHz speeds through the collective states with an emphasis on the low energy states that represent optimal solutions. As proof-of-concept, we present simulation results for standard NP-complete examples including a 16-city traveling salesman problem using experimentally benchmarked models for spin-transfer torque driven stochastic nanomagnets.

cond-mat.mes-hall↗

arXiv2017

Evaluating Spintronic Devices Using The Modular Approach

Over the past decade a large family of spintronic devices have been proposed as candidates for replacing CMOS for future digital logic circuits. Using the recently developed Modular Approach framework, we investigate and identify the physical bottlenecks and engineering challenges facing current spintronic devices. We then evaluate how systematic advancements in material properties and device design innovations impact the performance of spintronic devices, as a possible continuation of Moore's Law, even though some of these projections are speculative and may require technological breakthroughs. Lastly, we illustrate the use of the Modular Approach as an exploratory tool for probabilistic networks, using superparamagnetic magnets as building blocks for such networks. These building blocks leverage the inherent physics of stochastic spin-torque switching and could provide ultra-compact and efficient hardware for beyond-Boolean computational paradigms.

cond-mat.mes-hall↗

arXiv2016

Ultrafast Spin-Transfer-Torque Switching of Synthetic Ferrimagnets

The switching speed and the write current required for spin-transfer-torque reversal of spintronic devices such as magnetic tunnel junctions (MTJ) currently hinder their wide implementation into memory and logic devices. This problem is further exacerbated as the dimensions of MTJ nanostructures are scaled down to tens of nanometers in diameter, as higher magnetic anisotropy materials are required to meet thermal stability requirements that demand higher switching current densities. Here, we propose a simple solution to these issues based on synthetic ferrimagnet (SFM) structures. It is commonly assumed that to achieve a given switching delay, the current has to exceed the critical current by a certain factor and so a higher critical current implies a higher switching current. We show that this is not the case for SFM structures which can provide significantly reduced switching delay for a given current density, even though the critical current is increased. This non-intuitive result can be understood from the requirements of angular momentum conservation. We conclude that a 20 nm diameter MTJ incorporating the proposed SFM free layer structure can be switched in tens of picosecond time scales. This remarkable switching speed can be attained by employing current perpendicular magnetic anisotropy materials with experimentally demonstrated exchange coupling strengths.

cond-mat.mes-hall↗