Showing 1–10 of 10 results for author: Naeemi, A

Cross-layer Modeling and Design of Content Addressable Memories in Advanced Technology Nodes for Similarity Search

Authors: Siri Narla, Piyush Kumar, Mohammad Adnaan, Azad Naeemi

Abstract: In this paper we present a comprehensive design and benchmarking study of Content Addressable Memory (CAM) at the 7nm technology node in the context of similarity search applications. We design CAM cells based on SRAM, spin-orbit torque, and ferroelectric field effect transistor devices and from their layouts extract cell parasitics using state of the art EDA tools. These parasitics are used to de… ▽ More In this paper we present a comprehensive design and benchmarking study of Content Addressable Memory (CAM) at the 7nm technology node in the context of similarity search applications. We design CAM cells based on SRAM, spin-orbit torque, and ferroelectric field effect transistor devices and from their layouts extract cell parasitics using state of the art EDA tools. These parasitics are used to develop SPICE netlists to model search operations. We use a CAM-based dataset search and a sequential recommendation system to highlight the application-level performance degradation due to interconnect parasitics. We propose and evaluate two solutions to mitigate interconnect effects. △ Less

Submitted 22 March, 2024; originally announced March 2024.

Comments: 7 pages, 5 figures

Roadmap for Unconventional Computing with Nanotechnology

Authors: Giovanni Finocchio, Jean Anne C. Incorvia, Joseph S. Friedman, Qu Yang, Anna Giordano, Julie Grollier, Hyunsoo Yang, Florin Ciubotaru, Andrii Chumak, Azad J. Naeemi, Sorin D. Cotofana, Riccardo Tomasello, Christos Panagopoulos, Mario Carpentieri, Peng Lin, Gang Pan, J. Joshua Yang, Aida Todri-Sanial, Gabriele Boschetto, Kremena Makasheva, Vinod K. Sangwan, Amit Ranjan Trivedi, Mark C. Hersam, Kerem Y. Camsari, Peter L. McMahon , et al. (26 additional authors not shown)

Abstract: In the "Beyond Moore's Law" era, with increasing edge intelligence, domain-specific computing embracing unconventional approaches will become increasingly prevalent. At the same time, adopting a variety of nanotechnologies will offer benefits in energy cost, computational speed, reduced footprint, cyber resilience, and processing power. The time is ripe for a roadmap for unconventional computing w… ▽ More In the "Beyond Moore's Law" era, with increasing edge intelligence, domain-specific computing embracing unconventional approaches will become increasingly prevalent. At the same time, adopting a variety of nanotechnologies will offer benefits in energy cost, computational speed, reduced footprint, cyber resilience, and processing power. The time is ripe for a roadmap for unconventional computing with nanotechnologies to guide future research, and this collection aims to fill that need. The authors provide a comprehensive roadmap for neuromorphic computing using electron spins, memristive devices, two-dimensional nanomaterials, nanomagnets, and various dynamical systems. They also address other paradigms such as Ising machines, Bayesian inference engines, probabilistic computing with p-bits, processing in memory, quantum memories and algorithms, computing with skyrmions and spin waves, and brain-inspired computing for incremental learning and problem-solving in severely resource-constrained environments. These approaches have advantages over traditional Boolean computing based on von Neumann architecture. As the computational requirements for artificial intelligence grow 50 times faster than Moore's Law for electronics, more unconventional approaches to computing and signal processing will appear on the horizon, and this roadmap will help identify future needs and challenges. In a very fertile field, experts in the field aim to present some of the dominant and most promising technologies for unconventional computing that will be around for some time to come. Within a holistic approach, the goal is to provide pathways for solidifying the field and guiding future impactful discoveries. △ Less

Submitted 27 February, 2024; v1 submitted 17 January, 2023; originally announced January 2023.

Comments: 80 pages accepted in Nano Futures

Journal ref: Nano Futures (2024)

Physics-Based Models for Magneto-Electric Spin-Orbit Logic Circuits

Authors: Hai Li, Dmitri E. Nikonov, Chia-Ching Lin, Kerem Camsari, Yu-Ching Liao, Chia-Sheng Hsu, Azad Naeemi, Ian A. Young

Abstract: Spintronic devices are a promising beyond-CMOS device option thanks to their energy efficiency and compatibility with CMOS. To accurately capture their multi-physics dynamics, a rigorous treatment of both spin and charge and their inter-conversion is required. Here we present physics-based device models based on 4x4 matrices for the spin-orbit coupling part of the magneto-electric spin-orbit (MESO… ▽ More Spintronic devices are a promising beyond-CMOS device option thanks to their energy efficiency and compatibility with CMOS. To accurately capture their multi-physics dynamics, a rigorous treatment of both spin and charge and their inter-conversion is required. Here we present physics-based device models based on 4x4 matrices for the spin-orbit coupling part of the magneto-electric spin-orbit (MESO) device. Also, a more rigorous physics model of ferroelectric and magnetoelectric switching of ferromagnets, based on Landau-Lifshitz-Gilbert (LLG) and Landau-Khalatnikov (LK) equations, is presented. With the combined model implemented in a SPICE circuit simulator environment, simulation results were obtained which show feasibility of MESO implementation and functional operation of buffers, oscillators, and majority gates. △ Less

Submitted 21 October, 2021; originally announced October 2021.

Comments: 11 pages, 21 figures

Journal ref: IEEE Journal on Exploratory Solid-State Computational Devices and Circuits (2022)

A mixed signal architecture for convolutional neural networks

Authors: Qiuwen Lou, Chenyun Pan, John McGuiness, Andras Horvath, Azad Naeemi, Michael Niemier, X. Sharon Hu

Abstract: Deep neural network (DNN) accelerators with improved energy and delay are desirable for meeting the requirements of hardware targeted for IoT and edge computing systems. Convolutional neural networks (CoNNs) belong to one of the most popular types of DNN architectures. This paper presents the design and evaluation of an accelerator for CoNNs. The system-level architecture is based on mixed-signal,… ▽ More Deep neural network (DNN) accelerators with improved energy and delay are desirable for meeting the requirements of hardware targeted for IoT and edge computing systems. Convolutional neural networks (CoNNs) belong to one of the most popular types of DNN architectures. This paper presents the design and evaluation of an accelerator for CoNNs. The system-level architecture is based on mixed-signal, cellular neural networks (CeNNs). Specifically, we present (i) the implementation of different layers, including convolution, ReLU, and pooling, in a CoNN using CeNN, (ii) modified CoNN structures with CeNN-friendly layers to reduce computational overheads typically associated with a CoNN, (iii) a mixed-signal CeNN architecture that performs CoNN computations in the analog and mixed signal domain, and (iv) design space exploration that identifies what CeNN-based algorithm and architectural features fare best compared to existing algorithms and architectures when evaluated over common datasets -- MNIST and CIFAR-10. Notably, the proposed approach can lead to 8.7$\times$ improvements in energy-delay product (EDP) per digit classification for the MNIST dataset at iso-accuracy when compared with the state-of-the-art DNN engine, while our approach could offer 4.3$\times$ improvements in EDP when compared to other network implementations for the CIFAR-10 dataset. △ Less

Submitted 2 May, 2019; v1 submitted 30 October, 2018; originally announced November 2018.

Comments: 25 pages

Clocked Magnetostriction-Assisted Spintronic Device Design and Simulation

Authors: Rouhollah Mousavi Iraei, Nickvash Kani, Sourav Dutta, Dmitri E. Nikonov, Sasikanth Manipatruni, Ian A. Young, John T. Heron, Azad Naeemi

Abstract: We propose a heterostructure device comprised of magnets and piezoelectrics that significantly improves the delay and the energy dissipation of an all-spin logic (ASL) device. This paper studies and models the physics of the device, illustrates its operation, and benchmarks its performance using SPICE simulations. We show that the proposed device maintains low voltage operation, non-reciprocity, n… ▽ More We propose a heterostructure device comprised of magnets and piezoelectrics that significantly improves the delay and the energy dissipation of an all-spin logic (ASL) device. This paper studies and models the physics of the device, illustrates its operation, and benchmarks its performance using SPICE simulations. We show that the proposed device maintains low voltage operation, non-reciprocity, non-volatility, cascadability, and thermal reliability of the original ASL device. Moreover, by utilizing the deterministic switching of a magnet from the saddle point of the energy profile, the device is more efficient in terms of energy and delay and is robust to thermal fluctuations. The results of simulations show that compared to ASL devices, the proposed device achieves 21x shorter delay and 27x lower energy dissipation per bit for a 32-bit arithmetic-logic unit (ALU). △ Less

Submitted 22 November, 2017; originally announced November 2017.

Beyond-CMOS Device Benchmarking for Boolean and Non-Boolean Logic Applications

Authors: Chenyun Pan, Azad Naeemi

Abstract: The latest results of benchmarking research are presented for a variety of beyond-CMOS charge- and spin-based devices. In addition to improving the device-level models, several new device proposals and a few majorly modified devices are investigated. Deep pipelining circuits are employed to boost the throughput of low-power devices. Furthermore, the benchmarking methodology is extended to intercon… ▽ More The latest results of benchmarking research are presented for a variety of beyond-CMOS charge- and spin-based devices. In addition to improving the device-level models, several new device proposals and a few majorly modified devices are investigated. Deep pipelining circuits are employed to boost the throughput of low-power devices. Furthermore, the benchmarking methodology is extended to interconnect-centric analyses and non-Boolean logic applications. In contrast to Boolean circuits, non-Boolean circuits based on the cellular neural network demonstrate that spintronic devices can potentially outperform conventional CMOS devices. △ Less

Submitted 12 November, 2017; originally announced November 2017.

Comments: 9 pages, 10 figures

Non-volatile spin wave majority gate at the nanoscale

Authors: Odysseas Zografos, Sourav Dutta, Mauricio Manfrini, Adrien Vaysset, Bart Sorée, Azad Naeemi, Praveen Raghavan, Rudy Lauwereins, Iuliana P. Radu

Abstract: A spin wave majority fork-like structure with feature size of 40\,nm, is presented and investigated, through micromagnetic simulations. The structure consists of three merging out-of-plane magnetization spin wave buses and four magneto-electric cells serving as three inputs and an output. The information of the logic signals is encoded in the phase of the transmitted spin waves and subsequently st… ▽ More A spin wave majority fork-like structure with feature size of 40\,nm, is presented and investigated, through micromagnetic simulations. The structure consists of three merging out-of-plane magnetization spin wave buses and four magneto-electric cells serving as three inputs and an output. The information of the logic signals is encoded in the phase of the transmitted spin waves and subsequently stored as direction of magnetization of the magneto-electric cells upon detection. The minimum dimensions of the structure that produce an operational majority gate are identified. For all input combinations, the detection scheme employed manages to capture the majority phase result of the spin wave interference and ignore all reflection effects induced by the geometry of the structure. △ Less

Submitted 7 December, 2016; originally announced December 2016.

Journal ref: AIP Advances, Volume 7, Issue 5, 2017

Low-power Spin Valve Logic using Spin-transfer Torque with Automotion of Domain Walls

Authors: Sou-Chi Chang, Sasikanth Manipatruni, Dmitri E. Nikonov, Ian A. Young, Azad Naeemi

Abstract: A novel scheme for non-volatile digital computation is proposed using spin-transfer torque (STT) and automotion of magnetic domain walls (DWs). The basic computing element is composed of a lateral spin valve (SV) with two ferromagnetic (FM) wires served as interconnects, where DW automotion is used to propagate the information from one device to another. The non-reciprocity of both device and inte… ▽ More A novel scheme for non-volatile digital computation is proposed using spin-transfer torque (STT) and automotion of magnetic domain walls (DWs). The basic computing element is composed of a lateral spin valve (SV) with two ferromagnetic (FM) wires served as interconnects, where DW automotion is used to propagate the information from one device to another. The non-reciprocity of both device and interconnect is realized by sizing different contact areas at the input and the output as well as enhancing the local damping mechanism. The proposed logic is suitable for scaling due to a high energy barrier provided by a long FM wire. Compared to the scheme based on non-local spin valves (NLSVs) in the previous proposal, the devices can be operated at lower current density due to utilizing all injected spins for local magnetization reversals, and thus improve both energy efficiency and resistance to electromigration. This device concept is justified by simulating a buffer, an inverter, and a 3-input majority gate with comprehensive numerical simulations, including spin transport through the FM/non-magnetic (NM) interfaces as well as the NM channel and stochastic magnetization dynamics inside FM wires. In addition to digital computing, the proposed framework can also be used as a transducer between DWs and spin currents for higher wiring flexibility in the interconnect network. △ Less

Submitted 20 September, 2016; originally announced September 2016.

Comments: 9 pages

A Proposal for Energy-Efficient Cellular Neural Network based on Spintronic Devices

Authors: Chenyun Pan, Azad Naeemi

Abstract: Due to the massive parallel computing capability and outstanding image and signal processing performance, cellular neural network (CNN) is one promising type of non-Boolean computing system that can outperform the traditional digital logic computation and mitigate the physical scaling limit of the conventional CMOS technology. The CNN was originally implemented by VLSI analog technologies with ope… ▽ More Due to the massive parallel computing capability and outstanding image and signal processing performance, cellular neural network (CNN) is one promising type of non-Boolean computing system that can outperform the traditional digital logic computation and mitigate the physical scaling limit of the conventional CMOS technology. The CNN was originally implemented by VLSI analog technologies with operational amplifiers and operational transconductance amplifiers as neurons and synapses, respectively, which are power and area consuming. In this paper, we propose a hybrid structure to implement the CNN with magnetic components and CMOS peripherals with a complete driving and sensing circuitry. In addition, we propose a digitally programmable magnetic synapse that can achieve both positive and negative values of the templates. After rigorous performance analyses and comparisons, optimal energy is achieved based on various design parameters, including the driving voltage and the CMOS driving size. At a comparable footprint area and operation speed, a spintronic CNN is projected to achieve more than one order of magnitude energy reduction per operation compared to its CMOS counterpart. △ Less

Submitted 15 April, 2016; originally announced April 2016.

Smart Detector Cell: A Scalable All-Spin Circuit for Low Power Non-Boolean Pattern Recognition

Authors: Hamidreza Aghasi, Rouhollah Mousavi Iraei, Azad Naeemi, Ehsan Afshari

Abstract: We present a new circuit for non-Boolean recognition of binary images. Employing all-spin logic (ASL) devices, we design logic comparators and non-Boolean decision blocks for compact and efficient computation. By manipulation of fan-in number in different stages of the circuit, the structure can be extended for larger training sets or larger images. Operating based on the mainly similarity idea, t… ▽ More We present a new circuit for non-Boolean recognition of binary images. Employing all-spin logic (ASL) devices, we design logic comparators and non-Boolean decision blocks for compact and efficient computation. By manipulation of fan-in number in different stages of the circuit, the structure can be extended for larger training sets or larger images. Operating based on the mainly similarity idea, the system is capable of constructing a mean image and compare it with a separate input image within a short decision time. Taking advantage of the non-volatility of ASL devices, the proposed circuit is capable of hybrid memory/logic operation. Compared with existing CMOS pattern recognition circuits, this work achieves a smaller footprint, lower power consumption, faster decision time and a lower operational voltage. To the best of our knowledge, this is the first fully spin-based complete pattern recognition circuit demonstrated using spintronic devices. △ Less

Submitted 15 February, 2016; v1 submitted 10 May, 2015; originally announced May 2015.

Comments: This article is accepted to appear in IEEE Transactions on Nanotechnology