Stuck-at Fault Tolerance in RRAM Computing Systems
The DOI matches its Crossref record; 30 of the 31 references it cites were found in a registry or an index.
Abstract
Emerging metal-oxide resistive switching random-access memory (RRAM) devices and RRAM crossbars have demonstrated their potential in boosting the speed and energy-efficiency of analog matrix-vector multiplication. However, due to the immature fabrication technology, commonly occurring Stuck-At-Faults (SAFs) seriously degrade the computational accuracy of an RRAM-based computing system (RCS). In this paper, we present a fault-tolerant framework for RCS. A mapping algorithm with inner fault tolerance is proposed to convert matrix parameters into RRAM conductances in RCS and tolerate SAFs by fully exploring the available mapping space. Two baseline redundancy schemes are proposed to ensure that RCS is effective when the percentage of faulty RRAM cells is high. To reduce the number of redundant RRAM cells when the SAFs follow a non-uniform distribution or an unknown distribution, a distribution-aware redundancy scheme and a re-configurable redundancy scheme are proposed to provide dynamic fault tolerance. Simulation results show that, the baseline redundancy schemes can improve the recognition accuracy of the MNIST data set to almost the same as the RRAM-fault-free case, with an energy overhead of approximately 30%. When SAFs follow a non-uniform and an unknown distribution, the distribution-aware and re-configurable schemes can reduce the number of redundant RRAM cells from more than 200% to less than 40% and 60%, respectively, without reducing the recognition accuracy.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“However, grounding the bitlines through a resistor might alter the potential applied to the bitline, which will no longer be 0 volts, adding variability and thus altering the read over the sensing resistor100,112.”
From Hardware implementation of memristor-based artificial neural networks (2024) · cites this paper as [112]
Citations
Cites
Wong, H.-S. P., Lee, H.-Y., Yu, S., Chen, Y.-S., Wu, Y., Chen, P.-S., Lee, B., Chen, F. T., & Tsai, M.-J. (2012) · Proceedings of the IEEE · doi:10.1109/jproc.2012.2190369 · Cited by 2,757 · cited 1×
“RRAM is an emerging passive two-ports device based on metal oxide materials such as HfOx , TiOx and WOx [17], which has been widely used as non-volatile memory.”
BackgroundISAAC: a convolutional neural network accelerator with in-situ analog arithmetic in crossbars
Shafiee, A., Nag, A., Muralimanohar, N., Balasubramonian, R., Strachan, J. P., Hu, M., Williams, R. S., & Srikumar, V. (2016) · ACM SIGARCH Computer Architecture News · doi:10.1145/3007787.3001139 · Cited by 1,536 · cited 2×
“, RRAM-based dot-product operator [4], [20], RRAM-based fully-connected Neural Networks (NN) [21], [22], and RRAM-based convolutional neural networks [15], [23].”
MethodChi, P., Li, S., Xu, C., Zhang, T., Zhao, J., Liu, Y., Wang, Y., & Xie, Y. (2016) · ACM SIGARCH Computer Architecture News · doi:10.1145/3007787.3001140 · Cited by 1,231 · cited 2×
“The most attractive feature of RRAM devices is that the crossbar structure of RRAM array can be used to perform matrixvector multiplication, which efficiently merges the memory and computation together [18] and reduces the computation complexity from O(n2) to O(1) [19].”
Self-citationHu, M., Strachan, J. P., Li, Z., Grafals, E. M., Davila, N., Graves, C., Lam, S., Ge, N., Yang, J. J., & Williams, R. S. (2016) · Proceedings of the 53rd Annual Design Automation Conference · doi:10.1145/2897937.2898010 · Cited by 686 · cited 1×
“, RRAM-based dot-product operator [4], [20], RRAM-based fully-connected Neural Networks (NN) [21], [22], and RRAM-based convolutional neural networks [15], [23].”
Memristor Crossbar-Based Neuromorphic Computing System: A Case Study
Hu, M., Li, H., Chen, Y., Wu, Q., Rose, G. S., & Linderman, R. W. (2014) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2013.2296777 · Cited by 414 · cited 3×
“Previously proposed mapping algorithms, which convert matrix parameters into RRAM conductances of the RCS, use simplified assumptions and are targeted at ideal RRAM devices without SAFs [4], [13], [14].”
BackgroundMethodReal-time EEG analysis with subject-specific spatial patterns for a brain-computer interface (BCI)
Guger, C., Ramoser, H., & Pfurtscheller, G. (2000) · IEEE Transactions on Rehabilitation Engineering · doi:10.1109/86.895947 · Cited by 383 · cited 1×
“2776980 GPU, and FPGA, cannot meet the requirements of processing data in real-time with high energy efficiency [1].”
BackgroundRRAM Defect Modeling and Failure Analysis Based on March Test and a Novel Squeeze-Search Scheme
Chen, C.-Y., Shih, H.-C., Wu, C.-W., Lin, C.-H., Chiu, P.-F., Sheu, S.-S., & Chen, F. T. (2015) · IEEE Transactions on Computers · doi:10.1109/tc.2014.12 · Cited by 261 · cited 9×
“For example, researchers reported that only 63% of the RRAM cells are fault-free for a fabricated 4-Mb HfO2-based RRAM test chip, and SAFs appear rather frequently - about 10% of the RRAM cells in the chip contain SAFs [5].”
BackgroundMethodInfluentialRescuing Memristor-based Neuromorphic Design with High Defects
Liu, C., Hu, M., Strachan, J. P., & Li, H. (. (2017) · Proceedings of the 54th Annual Design Automation Conference 2017 · doi:10.1145/3061639.3062310 · Cited by 257 · cited 3×
“In future work, we will combine the proposed hardwarelevel schemes with algorithm-level methods [8], [9], [27] to further improve the fault tolerance of RCS.”
MethodAccelerator-friendly neural-network training: Learning variations and defects in RRAM crossbar
Chen, L., Li, J., Chen, Y., Deng, Q., Shen, J., Liang, X., & Jiang, L. (2017) · Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017 · doi:10.23919/date.2017.7926952 · Cited by 245 · cited 4×
“In future work, we will combine the proposed hardwarelevel schemes with algorithm-level methods [8], [9], [27] to further improve the fault tolerance of RCS.”
BackgroundMethodInfluentialTechnological Exploration of RRAM Crossbar Array for Matrix-Vector Multiplication
Xia, L., Gu, P., Li, B., Tang, T., Yin, X., Huangfu, W., Yu, S., Cao, Y., Wang, Y., & Yang, H. (2016) · Journal of Computer Science and Technology · doi:10.1007/s11390-016-1608-8 · Cited by 183 · cited 6×
“In prior work, after an RCS is fabricated, the parameters in the computed matrices are first mapped to the conductances of RRAM cells [4], [14].”
BackgroundMethodSelf-citationMNSIM: Simulation Platform for Memristor-based Neuromorphic Computing System
Xia, L., Li, B., Tang, T., Gu, P., Chen, P.-Y., Yu, S., Cao, Y., Wang, Y., Xie, Y., & Yang, H. (2017) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2017.2729466 · Cited by 171 · cited 1×
“, RRAM-based dot-product operator [4], [20], RRAM-based fully-connected Neural Networks (NN) [21], [22], and RRAM-based convolutional neural networks [15], [23].”
Self-citationi2WAP: Improving non-volatile cache lifetime by reducing inter- and intra-set write variations
Wang, J., Dong, X., Xie, Y., & Jouppi, N. P. (2013) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca.2013.6522322 · Cited by 152 · cited 4×
“Previous testing results have reported the distribution of SAFs for several types of RRAM devices [5], [12], and the distribution may be unknown for RRAM devices using emerging materials [24]–[26].”
BackgroundInfluentialSelf-citationFault-Tolerant Training with On-Line Fault Detection for RRAM-Based Neural Computing Systems
Xia, L., Liu, M., Ning, X., Chakrabarty, K., & Wang, Y. (2017) · Proceedings of the 54th Annual Design Automation Conference 2017 · doi:10.1145/3061639.3062248 · Cited by 152 · cited 2×
“In future work, we will combine the proposed hardwarelevel schemes with algorithm-level methods [8], [9], [27] to further improve the fault tolerance of RCS.”
BackgroundMethodSelf-citationCauses and consequences of the stochastic aspect of filamentary RRAM
Degraeve, R., Fantini, A., Raghavan, N., Goux, L., Clima, S., Govoreanu, B., Belmonte, A., Linten, D., & Jurczak, M. (2015) · Microelectronic Engineering · doi:10.1016/j.mee.2015.04.025 · Cited by 145 · cited 1×
“Since the metal-oxide layer of RRAM is sensitive to the fabrication thickness and the forming method [6], it is hard to prevent SAFs during fabrication process [7].”
MethodRENO: a high-efficient reconfigurable neuromorphic computing accelerator design
Liu, X., Mao, M., Liu, B., Li, H., Chen, Y., Li, B., Wang, Y., Jiang, H., Barnell, M., Wu, Q., et al. (2015) · Proceedings of the 52nd Annual Design Automation Conference · doi:10.1145/2744769.2744900 · Cited by 142 · cited 1×
“, RRAM-based dot-product operator [4], [20], RRAM-based fully-connected Neural Networks (NN) [21], [22], and RRAM-based convolutional neural networks [15], [23].”
Self-citationRRAM-Based Analog Approximate Computing
Li, B., Gu, P., Shan, Y., Wang, Y., Chen, Y., & Yang, H. (2015) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2015.2445741 · Cited by 135 · cited 1×
“The most attractive feature of RRAM devices is that the crossbar structure of RRAM array can be used to perform matrixvector multiplication, which efficiently merges the memory and computation together [18] and reduces the computation complexity from O(n2) to O(1) [19].”
MethodSelf-citationSwitched by input: power efficient structure for RRAM-based convolutional neural network
Xia, L., Tang, T., Huangfu, W., Cheng, M., Yin, X., Li, B., Wang, Y., & Yang, H. (2016) · Proceedings of the 53rd Annual Design Automation Conference · doi:10.1145/2897937.2898101 · Cited by 100 · cited 2×
“, RRAM-based dot-product operator [4], [20], RRAM-based fully-connected Neural Networks (NN) [21], [22], and RRAM-based convolutional neural networks [15], [23].”
MethodSelf-citationDesign Procedures for Three-Stage CMOS OTAs With Nested-Miller Compensation
Cannizzaro, S. O., Grasso, A. D., Mita, R., Palumbo, G., & Pennisi, S. (2007) · IEEE Transactions on Circuits and Systems I Fundamental Theory and Applications · doi:10.1109/tcsi.2007.895520 · Cited by 95 · cited 1×
“Without voltage followers or buffer amplifiers, the parameters of the neighboring layers in an RCS will influence each other through the sensing resistors [29], [30].”
BackgroundChen, Y. Y., Degraeve, R., Clima, S., Govoreanu, B., Goux, L., Fantini, A., Kar, G. S., Pourtois, G., Groeseneken, G., Wouters, D. J., et al. (2012) · Technical digest - International Electron Devices Meeting/Technical digest · doi:10.1109/iedm.2012.6479079 · Cited by 93 · cited 1×
“Since the metal-oxide layer of RRAM is sensitive to the fabrication thickness and the forming method [6], it is hard to prevent SAFs during fabrication process [7].”
MethodA CMOS Low-Dropout Regulator With Current-Mode Feedback Buffer Amplifier
Oh, W.-S., & Bakkaloğlu, B. (2007) · IEEE Transactions on Circuits and Systems II Analog and Digital Signal Processing · doi:10.1109/tcsii.2007.901621 · Cited by 86
MNSIM: Simulation Platform for Memristor-based Neuromorphic Computing System
Xia, L., Li, B., Tang, T., Gu, P., Yin, X., Huangfu, W., Chen, P.-Y., Yu, S., Cao, Y., Wang, Y., et al. (2016) · Proceedings of the 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE) · doi:10.3850/9783981537079_0549 · Cited by 75
Self-citationEnergy Efficient RRAM Spiking Neural Network for Real Time Classification
Wang, Y., Tang, T., Xia, L., Li, B., Gu, P., Yang, H., Li, H. H., & Xie, Y. (2015) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/2742060.2743756 · Cited by 53 · cited 2×
“The emerging metal-oxide resistive switching random-access memory (RRAM) has demonstrated tremendous potential in boosting the speed and energy efficiency of next-generation computing systems [2].”
BackgroundSelf-citationICE: inline calibration for memristor crossbar-based computing engine
Li, B., Wang, Y., Chen, Y., Li, H. H., & Yang, H. (2014) · doi:10.5555/2616606.2616832 · Cited by 43
Self-citationAn Efficient Algorithm for Spare Allocation Problems
Lin, H., Yeh, F.-M., & Kuo, S. (2006) · IEEE Transactions on Reliability · doi:10.1109/tr.2006.874942 · Cited by 38
Detection, diagnosis, and repair of faults in memristor-based memories
Kannan, S., Karimi, N., Karri, R., & Sinanoglu, O. (2014) · IEEE VLSI Test Symposium (VTS) · doi:10.1109/vts.2014.6818762 · Cited by 28 · cited 1×
“All existing RRAM-related redundancy schemes focus exclusively on memory applications, where the basic storage unit of an RRAM-based memory is a single RRAM cell [10]–[12].”
BackgroundICE: Inline calibration for memristor crossbar-based computing engine
Li, B., Wang, Y., Chen, Y., Li, H. H., & Yang, H. (2014) · Design, Automation, and Test in Europe · doi:10.7873/date2014.197 · Cited by 24
Self-citationPhysical mechanism of HfO2-based bipolar resistive random access memory
Chang, H. L., Li, H.-C., Liu, C. W., Chen, F., & Tsai, M.-J. (2011) · International Symposium on VLSI Technology, Systems and Applications (VLSI-TSA) · doi:10.1109/vtsa.2011.5872253 · Cited by 22 · cited 1×
“Previous testing results have reported the distribution of SAFs for several types of RRAM devices [5], [12], and the distribution may be unknown for RRAM devices using emerging materials [24]–[26].”
BackgroundMapping neural network computations onto memristor crossbar
Tarkov, M. S. (2015) · International Siberian Conference on Control and Communications · doi:10.1109/sibcon.2015.7147235 · Cited by 18 · cited 1×
“Previously proposed mapping algorithms, which convert matrix parameters into RRAM conductances of the RCS, use simplified assumptions and are targeted at ideal RRAM devices without SAFs [4], [13], [14].”
MethodTraining-based forming process for RRAM yield improvement
Shih, H.-C., Chen, C., Wu, C., Lin, C.-H., & Sheu, S.-S. (2011) · IEEE VLSI Test Symposium (VTS) · doi:10.1109/vts.2011.5783775 · Cited by 11 · cited 1×
“Although some techniques have been proposed to repair the cells with SAFs post-fabrication [28], there will still be many unrecoverable devices.”
BackgroundResisitive switching variability study on 1T1R AlOx/WOx-based RRAM array
Jiao, B., Deng, N., Yu, J., Bai, Y., Wu, M., Ye, Z., Qian, H., & Wu, H. (2013) · IEEE International Conference on Electron Devices and Solid-State Circuits (EDSSC) · doi:10.1109/edssc.2013.6628071 · Cited by 7 · cited 2×
“In some RRAM devices using emerging material and fabrication technologies, although the yield of RRAM cells has been reported using single-cell testing [25], [26], the distribution of SAFs in an RRAM crossbar may not be easy to predict before fabrication.”
Method
Research works printed but not linked: 2
- lecun, 0, The MNIST Database of Handwritten Digits
- prakash, 2013, Improved resistance memory characteristics and switching mechanism using TiN electrode on TaOx/W structure, Proc Winter Simulation Conf, 136
Cited by
Hardware implementation of memristor-based artificial neural networks
Aguirre, F., Sebastian, A., Le Gallo, M., Song, W., Wang, T., Yang, J. J., Lu, W., Chang, M.-F., Ielmini, D., Yang, Y., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-45670-9 · Cited by 564
“There are three widely used sensing modes for the output voltages 112 .”
Reliability of analog resistive switching memory for neuromorphic computing
Zhao, M., Gao, B., Tang, J., Qian, H., & Wu, H. (2020) · Applied Physics Reviews · doi:10.1063/1.5124915 · Cited by 349
Mehonic, A., Sebastian, A., Rajendran, B., Simeone, O., Vasilaki, E., & Kenyon, A. J. (2020) · Advanced Intelligent Systems · doi:10.1002/aisy.202000085 · Cited by 331
Committee machines—a universal method to deal with non-idealities in memristor-based neural networks
Joksas, D., Freitas, P., Chai, Z., Ng, W. H., Buckwell, M., Li, C., Zhang, W., Xia, Q., Kenyon, A. J., & Mehonić, A. (2020) · RePEc: Research Papers in Economics · Cited by 237
He, Z., Lin, J., Ewetz, R., Yuan, J.-S., & Fan, D. (2019) · Proceedings of the 56th Annual Design Automation Conference 2019 · doi:10.1145/3316781.3317870 · Cited by 149
“Many previous studies [3, 17] only discuss the effect of SAF on NN inference, using a simple 2-layers fully-connected network on MNIST, where the results show that SAF indeed affects the accuracy but with a relatively low degradation.”
BackgroundMethod
Metadata
- Authors
- Xia, L., Huangfu, W., Tang, T., Yin, X., Chakrabarty, K., Xie, Y., Wang, Y., & Yang, H.
- Affiliations
- Tsinghua University (CN)
- University of California, Santa Barbara (US)
- Duke University (US)
- Venue
- IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Published
- 2018-03
- Volume
- 8
- Issue
- 1
- Pages
- 102-115
- ISSN
- 2156-3357 (ISSN-L)
- 2156-3365
- OpenAlex
- W2768104155
- Cited by
- 177
- References
- 31
- Citations by year
- 2026: 17 · 2025: 23 · 2024: 25 · 2023: 18 · 2022: 33 · 2021: 20 · 2020: 23 · 2019: 13 · 2018: 5
- FWCI
- 4.66
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Journal on Emerging and Selected Topics in Circuits and Systems (publishedVersion)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- Funders
- National Key Research and Development Program of China (2017YFA0207600)
- National Natural Science Foundation of China (61373026, 61622403, 61621091)
- Joint Fund of Equipment Pre-Research and Ministry of Education (6141A02022608)
- U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research (DE-SC0013553)
- NSF (1461698, 1500848)
- Qualcomm
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
- Physical Sciences › Computer Science › Hardware and Architecture › VLSI and Analog Circuit Testing
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Ferroelectric and Negative Capacitance Devices
- Keywords
- stuck-at faults, fault tolerance, resistive random access memory, redundancy schemes, in-memory computing, energy overhead, MNIST recognition
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.