Training-based forming process for RRAM yield improvement

Shih, Hsiu-Chuan; Chen, Ching-Yi; Wu, Cheng‐Wen; Lin, Chih-Hsiou; Sheu, Shyh-Shyuan

doi:10.1109/vts.2011.5783775

Cited by 9 publications

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The existence of stuck-on and stuckoff devices can both affect the accuracy of a neural network, but the influence of a stuck-on device is much worse [81]. Shih et al improved the yield based on a training sequence to solve the over forming problem [82]. The accuracy degradation caused by a stuck device can be solved by optimizing the weight mapping method, such as introducing redundancy RRAM rows [83] or remapping the synaptic weight considering the distribution of stuck devices [84].…”

Section: Yieldmentioning

confidence: 99%

Oxide-based filamentary RRAM for deep learning

Zhang¹,

Huang²,

Gao³

et al. 2020

J. Phys. D: Appl. Phys.

View full text Add to dashboard Cite

We provide an overview of the field of oxide-based filamentary resistive random access memory (RRAM) for deep learning neural networks (DNNs). After introducing the electrical characteristics of filamentary RRAM which are the main influences on the performance of DNNs, we show several cases focusing on the optimization of the linearity and symmetry of conductance modulation and precision. Progress in the optimization of energy efficiency and reliability of DNNs is then reviewed. Lastly, our recent work on DNNs based on analog filamentary RRAM, including the design and fabrication of the device, is introduced in detail.

show abstract

Section: Yieldmentioning

confidence: 99%

Oxide-based filamentary RRAM for deep learning

Zhang¹,

Huang²,

Gao³

et al. 2020

J. Phys. D: Appl. Phys.

View full text Add to dashboard Cite

show abstract

Electrical characterization and modeling of pulse-based forming techniques in RRAM arrays

Grossi

Zambelli

Olivo

et al. 2016

Solid-State Electronics

View full text Add to dashboard Cite

The forming process, which corresponds to the activation of the switching filament in Resistive Random Access Memory (RRAM) arrays, has a strong impact on the cells' performances. In this paper we characterize and compare different pulse forming techniques in terms of forming time, yield and cell-to-cell variability on 4 kbits RRAM arrays. Moreover, post-forming modeling during Reset operation of correctly working and over formed cells has been performed. An incremental form and verify technique, based on a sequence of trapezoidal waveforms with increasing voltages followed by a verify operation that terminates when the expected switching behavior has been achieved, showed the best results. This procedure narrows the post-forming current distribution whereas reducing the reset switching voltage and the operative current. These advantages materialize in a better control of the cell-to-cell variability and in an overall time and energy saving at the system level.

show abstract

Back-End-of-Line Defect Analysis for Rnv8T Nonvolatile SRAM

Bai

Hsu

et al. 2013

2013 22nd Asian Test Symposium

View full text Add to dashboard Cite

Training-based forming process for RRAM yield improvement

Cited by 9 publications

References 16 publications

Oxide-based filamentary RRAM for deep learning

Oxide-based filamentary RRAM for deep learning

Electrical characterization and modeling of pulse-based forming techniques in RRAM arrays

Back-End-of-Line Defect Analysis for Rnv8T Nonvolatile SRAM

Contact Info

Product

Resources

About