Effect of Device Variation on Mapping Binary Neural Network to Memristor Crossbar Array

Yi, Won Ju; Kim, Yulhwa; Kim, Jae‐Joon

doi:10.23919/date.2019.8714817

Cited by 10 publications

(11 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The intrinsic stochastic nature of the switching mechanism can induce a high degree of variability [92]. Such variability in the memristive structures also increases the possibility of a deviation of the weighted sum from the target conductance value [93]. The variability of the conductance states G LRS and G HRS across a matrix is largely influenced by the choice of the stack's materials (e.g., single material HfO X vs. bilayer HfO X + TaO X ) [94,95], as well as by the device scaling.…”

Section: Frequency and Set Transition Abruptnessmentioning

confidence: 99%

See 1 more Smart Citation

SPICE Simulation of RRAM-Based Cross-Point Arrays Using the Dynamic Memdiode Model

et al. 2021

View full text Add to dashboard Cite

We thoroughly investigate the performance of the Dynamic Memdiode Model (DMM) when used for simulating the synaptic weights in large RRAM-based cross-point arrays (CPA) intended for neuromorphic computing. The DMM is in line with Prof. Chua’s memristive devices theory, in which the hysteresis phenomenon in electroformed metal-insulator-metal structures is represented by means of two coupled equations: one equation for the current-voltage characteristic of the device based on an extension of the quantum point-contact (QPC) model for dielectric breakdown and a second equation for the memory state, responsible for keeping track of the previous history of the device. By considering ex-situ training of the CPA aimed at classifying the handwritten characters of the MNIST database, we evaluate the performance of a Write-Verify iterative scheme for setting the crosspoint conductances to their target values. The total programming time, the programming error, and the inference accuracy obtained with such writing scheme are investigated in depth. The role played by parasitic components such as the line resistance as well as some CPA’s particular features like the dynamical range of the memdiodes are discussed. The interrelationship between the frequency and amplitude values of the write pulses is explored in detail. In addition, the effect of the resistance shift for the case of a CPA programmed with no errors is studied for a variety of input signals, providing a design guideline for selecting the appropriate pulse’s amplitude and frequency.

show abstract

Section: Frequency and Set Transition Abruptnessmentioning

confidence: 99%

“…Extreme scaling seems to reduce variability likely because of a reduction of the area where the switching occurs [96]. Furthermore, the conductance value set during the CLT procedure also exhibits variability [19,93].…”

Section: Frequency and Set Transition Abruptnessmentioning

confidence: 99%

SPICE Simulation of RRAM-Based Cross-Point Arrays Using the Dynamic Memdiode Model

et al. 2021

View full text Add to dashboard Cite

show abstract

“…However, the uncontrolled high density of defects in addition to the intrinsic stochastic nature of the switching mechanism can induce a high degree of variability [89]. Along with the read margin reduction caused by the limited resistance window (R ON /R OF F ) and interconnection line resistance (R W ), such variability in the memristive structures also increases the possibility of a deviation of the weighted sum from the target value [90]. Moreover, if each device performs slightly different and its characteristics are allowed to vary in time, programming to a desired state becomes a challenging task.…”

Section: Device-to-device and Programming Variabilitymentioning

confidence: 99%

“…Extreme scaling seems to reduce the variability, probably because of a reduction of the area where the switching occurs [93]. Furthermore, the conductance value set during the Write-Verify procedure also presents variability [56], [90]. As in this paper we exclusively focus on the inference phase, in this sub-Section we show that the QMM is suitable to study the device variability in terms of the R ON and R OF F dispersion, as well as due to errors in the conductance programming, expressed by the variability of the QMM control parameter λ (see Fig.…”

Section: Device-to-device and Programming Variabilitymentioning

confidence: 99%

See 1 more Smart Citation

Application of the Quasi-Static Memdiode Model in Cross-Point Arrays for Large Dataset Pattern Recognition

et al. 2020

View full text Add to dashboard Cite

We investigate the use and performance of the quasi-static memdiode model (QMM) when incorporated into large cross-point arrays intended for pattern classification tasks. Following Chua's memristive devices theory, the QMM comprises two equations, one equation for the electron transport based on the doublediode circuit with single series resistance and a second equation for the internal memory state of the device based on the so-called logistic hysteron or memory map. Ex-situ trained memdiodes with different MNISTlike databases are used to establish the synaptic weights linking the top and bottom wire networks. The role played by the memdiode electrical parameters, wire resistance and capacitance values, image pixelation, connection schemes, signal-to-noise ratio and device-to-device variability in the classification effectiveness are investigated. The confusion matrix is used to benchmark the system performance metrics. We show that the simplicity, accuracy and robustness of the memdiode model makes it a suitable candidate for the realistic simulation of large-scale neural networks with non-idealities.

show abstract

Simulation of the effect of material properties on yttrium oxide memristor-based artificial neural networks

Aguirre

Piros

Kaiser

et al. 2023

APL Machine Learning

View full text Add to dashboard Cite

This paper reports a simulation study concerning the effect of yttrium oxide stoichiometry on output features of a memristor-based single layer perceptron neural network. To carry out this investigation, a material-oriented behavioral compact model for bipolar-type memristive devices was developed and tested. The model is written for the SPICE (Simulation Program with Integrated Circuits Emphasis) simulator and considers as one of its inputs a measure of the oxygen flow used during the deposition of the switching layer. After a thorough statistical calibration of the model parameters using experimental current–voltage characteristics associated with different fabrication conditions, the corresponding curves were simulated and the results were compared with the original data. In this way, the average switching behavior of the structures (low and high current states, set and reset voltages, etc.) as a function of the oxygen content can be forecasted. In a subsequent phase, the collective response of the devices when used in a neural network was investigated in terms of the output features of the network (mainly power dissipation and power efficiency). The role played by parasitic elements, such as the line resistance and the read voltage influence on the inference accuracy, was also explored. Since a similar strategy can be applied to any other material-related fabrication parameter, the proposed approach opens up a new dimension for circuit designers, as the behavior of complex circuits employing devices with specific characteristics can be realistically assessed before fabrication.

show abstract

Effect of Device Variation on Mapping Binary Neural Network to Memristor Crossbar Array

Cited by 10 publications

References 8 publications

SPICE Simulation of RRAM-Based Cross-Point Arrays Using the Dynamic Memdiode Model

SPICE Simulation of RRAM-Based Cross-Point Arrays Using the Dynamic Memdiode Model

Application of the Quasi-Static Memdiode Model in Cross-Point Arrays for Large Dataset Pattern Recognition

Simulation of the effect of material properties on yttrium oxide memristor-based artificial neural networks

Contact Info

Product

Resources

About