Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks
The DOI matches its Crossref record; 38 of the 42 references it cites were found in a registry or an index.
Abstract
Eyeriss is an accelerator for state-of-the-art deep convolutional neural networks (CNNs). It optimizes for the energy efficiency of the entire system, including the accelerator chip and off-chip DRAM, for various CNN shapes by reconfiguring the architecture. CNNs are widely used in modern AI systems but also bring challenges on throughput and energy efficiency to the underlying hardware. This is because its computation requires a large amount of data, creating significant data movement from on-chip and off-chip that is more energy-consuming than computation. Minimizing data movement energy cost for any CNN shape, therefore, is the key to high throughput and energy efficiency. Eyeriss achieves these goals by using a proposed processing dataflow, called row stationary (RS), on a spatial architecture with 168 processing elements. RS dataflow reconfigures the computation mapping of a given shape, which optimizes energy efficiency by maximally reusing data locally to reduce expensive data movement, such as DRAM accesses. Compression and data gating are also applied to further improve energy efficiency. Eyeriss processes the convolutional layers at 35 frames/s and 0.0029 DRAM access/multiply and accumulation (MAC) for AlexNet at 278 mW (batch size N = 4), and 0.7 frames/s and 0.0035 DRAM access/MAC for VGG-16 at 236 mW (N = 3).
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
“such scaling will enable NeuRRAM to deliver higher energy and area efficiency than today’s state-of-the-art edge inference accelerators58–61.”
From A compute-in-memory chip based on resistive random-access memory (2022) · cites this paper as [58]
In the citing paper
“Some studies proposed to implement the argmax() function via hardware128–148, which could be beneficial to reduce the total transistor count and power consumption while at the same time increasing the throughput.”
From Hardware implementation of memristor-based artificial neural networks (2024) · cites this paper as [131]
In the citing paper
“Based on the analysis of the research, CoNNA is able to speed up at least 3.08 times and up to 14.1 times faster than the MIT Eyeriss Accelerator [20].”
From Embedded Intelligence on FPGA: Survey, Applications and Challenges (2021) · cites this paper as [20]
In the citing paper
“massively parallel processing architectures with distributed memory architectures, such as graphic processing units (GPUs) but also increasingly specialized chips for NN inference (e.g., [17]), have shown to be very energy efficient and capable of impressive computational throughput when fully utilized by concurrent algorithms [18].”
From An Introduction to Deep Learning for the Physical Layer (2017) · cites this paper as [17]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“Modern CNNs achieve their superior accuracy by building a very deep hierarchy of layers [2]–[5], [7], which transform the input image data into highly abstract representations called feature maps (fmaps).”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 7×
“We benchmark the chip performance using two publicly available and widely used CNNs: AlexNet [2] and VGG-16 [3].”
MethodInfluentialImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 1×
“[25] demonstrate a CNN processor and only report the theoretical peak throughput along with the power measured on a CNN for the MNIST data set [26], which has storage and computation requirements that are orders of magnitude lower than the state-of-the-art CNNs.”
MethodGoing deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380 · cited 2×
“We have demonstrated an 1000-class image classification task [27] using this system, and a live demo can be found in [29].”
BackgroundMethodRich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684
Mastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016) · Nature · doi:10.1038/nature16961 · Cited by 15,612 · cited 1×
“D EEP learning using convolutional neural networks (CNNs) [1] has achieved unprecedented accuracy on many modern AI applications [2]–[9].”
Background(2010) · International Conference on Machine Learning · Cited by 13,995
Research works printed but not linked: 2
- han, 2015, Learning both weights and connections for efficient neural network, Proc Adv Neural Inf Process Syst, 28, 1135
- chen, 2016, An 1000-Class Image Classification Task Performed by the Eyeriss-Integrated Deep Learning System
Cited by
Items 601–700 of 1,001
Wire-Aware Architecture and Dataflow for CNN Accelerators
Gudaparthi, S., Narayanan, S., Balasubramonian, R., Giacomin, E., Hari, K., & Gaillardon, P. (2019) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1145/3352460.3358316 · Cited by 40
“For example, the Eyeriss architecture and its rowstationary dataflow require non-trivial storage for scratchpads and registers per processing element (PE) tomaximize reuse [11].”
BackgroundXu, W., Zhang, Z., You, X., & Zhang, C. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.2973355 · Cited by 32
“Comparison of processing time on AlexNet with Eyeriss [10] and DNA [11].”
BackgroundMethodInfluentialDNNGuard: An Elastic Heterogeneous DNN Accelerator Architecture against Adversarial Attacks
Wang, X., Hou, R., Zhao, B., Yuan, F., Zhang, J., Meng, D., & Qian, X. (2020) · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · doi:10.1145/3373376.3378532 · Cited by 35
“In contrast, for inference tasks that execute in resource-constraint environments, DNN accelerators are widely and commercially deployed [22, 23].”
BackgroundMethodMauro, A. D., Conti, F., Schiavone, P. D., Rossi, D., & Benini, L. (2020) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2020.3012576 · Cited by 42
PCNN: Pattern-based Fine-Grained Regular Pruning Towards Optimizing CNN Accelerators
Tan, Z., Song, J., Ma, X., Tan, S.-H., Chen, H., Miao, Y., Wu, Y., Ye, S., Wang, Y., Li, D., et al. (2020) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac18072.2020.9218498 · Cited by 36
“When neural networks achieve higher accuracy with increasing computation and parameters [4] [5], many accelerators [6] [7] [8] [9] implemented on ASIC or FPGA with abundant parallel units are proposed to deploy neural networks on edge devices.”
MethodPaliy, M., Strangio, S., Ruiu, P., Rizzo, T., & Iannaccone, G. (2020) · IEEE Access · doi:10.1109/access.2020.3037017 · Cited by 35
RecPipe: Co-designing Models and Hardware to Jointly Optimize Recommendation Quality and Performance
Gupta, U., Hsia, S., Zhang, J. J., Wilkening, M., Pombra, J., Lee, H.-H. S., Wei, G.-Y., Wu, C.-J., & Brooks, D. M. (2021) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1145/3466752.3480127 · Cited by 38
“Lots of research effort has been devoted to design specialized hardware for deep learning—especially MLPs, CNNs, and RNNs [4,6,7, 8,17,19,30,42,43,44,47,54,55].”
BackgroundMethodJie-Fang, Z., & Zhang, Z. (2021) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1145/3466752.3480081 · Cited by 31
“DNN/CNN accelerators [16, 8, 28, 38] are incapable of performing such computation because they are purposely designed for sequential memory access and regularstructured data, i.”
BackgroundValue-aware Parity Insertion ECC for Fault-tolerant Deep Neural Network
Lee, S.-S., & Yang, J.-S. (2022) · Design, Automation, and Test in Europe · doi:10.23919/date54114.2022.9774543 · Cited by 31
Hybrid Accumulator Factored Systolic Array for Machine Learning Acceleration
Inayat, K., & Chung, J. (2022) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2022.3170233 · Cited by 34
Trainer: An Energy-Efficient Edge-Device Training Processor Supporting Dynamic Weight Pruning
Wang, Y., Qin, Y., Deng, D., Wei, J., Chen, T., Lin, X., Liu, L., Wei, S., & Yin, S. (2022) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2022.3174411 · Cited by 31
Cao, N., Chang, M., & Raychowdhury, A. (2019) · Digest of technical papers/Digest of technical papers - IEEE International Solid-State Circuits Conference · doi:10.1109/isscc.2019.8662311 · Cited by 31
Advancing TinyML in IoT: A Holistic System-Level Perspective for Resource-Constrained AI
Ortiz, L. A. P., Soliz, I. F. M., & Balarezo, V. K. G. (2025) · Future Internet · doi:10.3390/fi17060257 · Cited by 43
“Demand for On-Device Intelligence: Applications like real-time anomaly detection in industrial machinery, immediate wake-word recognition in voice assistants, or continuous health monitoring on wearables require decisions to be made locally with minimal latency [4,17,18].”
Normalization and dropout for stochastic computing-based deep convolutional neural networks
Li, J., Yuan, Z., Li, Z., Ren, A., Ding, C., Draper, J. T., Nazarian, S., Qiu, Q., Yuan, B., & Wang, Y. (2017) · Integration · doi:10.1016/j.vlsi.2017.11.002 · Cited by 33
Kim, Y., Kim, H., Ahn, D., & Kim, J. (2018) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1145/3218603.3218605 · Cited by 33
“Therefore, various dedicated hardware accelerators for CNN applications have been introduced [4, 5, 7, 13].”
BackgroundLi, Z., Wang, J., Sylvester, D., Blaauw, D. T., & Kim, H.-S. (2019) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2018.2885559 · Cited by 30
Cavigelli, L., & Benini, L. (2019) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2019.2903421 · Cited by 32
Optimizing Performance of Convolutional Neural Network Using Computing Technique
Samudre, P., Shende, P., & Jaiswal, V. K. (2019) · IEEE International Conference for Convergence in Technology (I2CT) · doi:10.1109/i2ct45611.2019.9033876 · Cited by 33
“FPGA based accelerators have gained more interest of researchers because they have benefits of better result like high energy performance, quick development round and ability of reconfiguration [6].”
BackgroundYuan, Z., Yang, Y., Yue, J., Liu, R., Feng, X., Lin, Z., Wu, X., Li, X., Yang, H., & Liu, Y. (2020) · Digest of technical papers/Digest of technical papers - IEEE International Solid-State Circuits Conference · doi:10.1109/isscc19947.2020.9063155 · Cited by 30
“Existing CNN processors [1-6] are mainly aimed at single-image processing, ignoring activation similarity among sequential frames.”
BackgroundA 2.17-mW Acoustic DSP Processor With CNN-FFT Accelerators for Intelligent Hearing Assistive Devices
Lee, Y.-C., Chi, T.-S., & Yang, C. (2020) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2020.2987695 · Cited by 38
Recent Progress on Memristive Convolutional Neural Networks for Edge Intelligence
Qin, Y., Bao, H., Wang, F., Chen, J., Li, Y., & Miao, X. (2020) · Advanced Intelligent Systems · doi:10.1002/aisy.202000114 · Cited by 40
CRIME: Input-Dependent Collaborative Inference for Recurrent Neural Networks
Pagliari, D. J., Chiaro, R., Macii, E., & Poncino, M. (2020) · IEEE Transactions on Computers · doi:10.1109/tc.2020.3021199 · Cited by 36
“Many studies have focused on hardware accelerator designs, which exploit the parallelism of the matrix multiplication kernels that dominate Deep Neural Networks (DNN) inference, and leverage techniques such as weight quantization and pruning to reduce the memory bottleneck [12].”
BackgroundLow-Latency In Situ Image Analytics With FPGA-Based Quantized Convolutional Neural Network
Wang, M., Lee, K. C. M., Chung, B. M. F., Varma, B. S., Ng, H.-C., Wong, J. S. J., Shum, H. C., Tsia, K. K., & So, H. K. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3046452 · Cited by 33
Practically Deploying Heavyweight Adaptive Bitrate Algorithms With Teacher-Student Learning
Meng, Z., Guo, Y., Shen, Y., Chen, J., Zhou, C., Wang, M., Zhang, J., Xu, M., Sun, C., & Hu, H. (2021) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2020.3048666 · Cited by 31
Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication
Moon, G. E., Kwon, H., Jeong, G., Chatarasi, P., Rajamanickam, S., & Krishna, T. (2021) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2021.3104240 · Cited by 35
Self-citationSoltaniyeh, M., Martin, R. P., & Nagarakatte, S. (2022) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3532863 · Cited by 38
“6 RELATED WORK There is a large body of literature on using custom hardware accelerators to improve the performance and energy efficiency of neural networks [2, 5, 7, 9, 12, 14, 16, 31, 33, 34, 38, 42, 44].”
BackgroundMethodResultInfluentialChoi, W., Jeong, K., Choi, K., Lee, K., & Park, J. (2018) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac.2018.8465903 · Cited by 37
Sk, M. R., Thunder, S., Lehninger, D., Sanctis, S., Raffel, Y., Lederer, M., Jank, M. P. M., Kämpfe, T., De, S., & Chakrabarti, B. (2023) · ACS Applied Electronic Materials · doi:10.1021/acsaelm.2c01357 · Cited by 32
Liquid-Based Memory Devices for Next-Generation Computing
Kim, D., & Lee, J. (2023) · ACS Applied Electronic Materials · doi:10.1021/acsaelm.2c01636 · Cited by 30
Perspective: Uniform switching of artificial synapses for large-scale neuromorphic arrays
Tan, S. H., Lin, P., Yeon, H., Choi, S., Park, Y., & Kim, J. (2018) · APL Materials · doi:10.1063/1.5049137 · Cited by 33
Neural networks on microcontrollers: saving memory at inference via operator reorderingPreprint
Liberis, E., & Lane, N. D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.05110 · Cited by 66
“Concerns about the memory usage of neural networks and data movement during execution are also discussed in neural network accelerator chip design literature [29–31].”
BackgroundAn Energy-and-Area-Efficient CNN Accelerator for Universal Powers-of-Two Quantization
Xia, T., Zhao, B., Ma, J., Fu, G., Zhao, W., Zheng, N., & Ren, P. (2022) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2022.3227608 · Cited by 39
YOLOv8-Lite: A Lightweight Object Detection Model for Real-time Autonomous Driving Systems
Yang, M., & Fan, X. (2024) · ICCK Transactions on Emerging Topics in Artificial Intelligence · doi:10.62762/tetai.2024.894227 · Cited by 47
Lin, Y., Zhang, S., & Shanbhag, N. R. (2016) · SiPS ... design and implementation - IEEE Workshop on Signal Processing Systems · doi:10.1109/sips.2016.11 · Cited by 32
“Simulation results are shown in Section V. Finally, conclusions and future work are presented in Section VI.”
BackgroundMethodDynamic Bit-width Reconfiguration for Energy-Efficient Deep Learning Hardware
Pagliari, D. J., Macii, E., & Poncino, M. (2018) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1145/3218603.3218611 · Cited by 38
“Several works have proposed implementations of low-power HW accelerators for CNNs on FPGAs [20] and ASICs [4, 7, 14, 16, 21].”
BackgroundAn Analytical Method to Determine Minimum Per-Layer Precision of Deep Neural Networks
Sakr, C., & Shanbhag, N. R. (2018) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2018.8461702 · Cited by 44
“Among recent approaches are pruning [4], sequential reduction [5], and zero skipping [1].”
BackgroundMethodQin, Q., Ren, J., Yu, J., Wang, H., Gao, L. C., Zheng, J., Feng, Y., Fang, J., & Wang, Z. (2018) · IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA) · doi:10.1109/bdcloud.2018.00110 · Cited by 39
Efficient Modular Adder Designs Based on Thermometer and One-Hot Coding
Jafarzadehpour, F., Molahosseini, A. S., Zarandi, A. A. E., & Sousa, L. A. (2019) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2019.2919609 · Cited by 29
A 1-Mbit Fully Logic-Compatible 3T Gain-Cell Embedded DRAM in 16-nm FinFET
Giterman, R., Shalom, A., Burg, A. P., Fish, A., & Teman, A. (2020) · IEEE Solid-State Circuits Letters · doi:10.1109/lssc.2020.3006496 · Cited by 35
“While 6T SRAM has been the topology of choice for most systems, its size and poor energy efficiency often limit the cost and performance of emerging applications such as machine learning accelerators [1], [2].”
BackgroundAn Edge 3D CNN Accelerator for Low-Power Activity Recognition
Wang, Y., Wang, Y., Shi, C., Cheng, L., Li, H., & Li, X. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.3011042 · Cited by 30
“We also use the same analytical model proposed in [14] to perform an overall energy estimation of different implementations.”
BackgroundMethodAn Elastic Task Scheduling Scheme on Coarse-Grained Reconfigurable Architectures
Chen, L., Zhu, J., Deng, Y. S., Li, Z., Chen, J., Jiang, X., Yin, S., Wei, S., & Liu, L. (2021) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2021.3084804 · Cited by 29
“RECENT years have witnessed strong momentum in exploiting hardware accelerators for a higher level of performance and energy efficiency [1], [2], [3], [4], [5], [6].”
BackgroundESCALATE: Boosting the Efficiency of Sparse CNN Accelerator with Kernel Decomposition
Li, S., Hanson, E., Qian, X., Li, H., & Chen, Y. (2021) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1145/3466752.3480043 · Cited by 31
“Dense accelerators [5–7] optimize the inference efficiency by exploring the reuse opportunities and designing efficient dataflows.”
BackgroundMethodSelf-citationNeural Network Training With Asymmetric Crosspoint Elements
Onen, M., Gokmen, T., Todorov, T. K., Nowicki, T., del Alamo, J. A., Rozen, J., Hänsch, W., & Kim, S. (2022) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2022.891624 · Cited by 36
“As a result, the necessity to provide fast and energy-efficient solutions for deep learning has invoked a massive collective research effort by industry and academia (Chen et al., 2016; Jouppi et al., 2017; Rajbhandari et al., 2020).”
BackgrounduSystolic: Byte-Crawling Unary Systolic Array
Wu, D., & Miguel, J. S. (2022) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca53966.2022.00010 · Cited by 37
“The edge and cloud configurations take their shapes from MIT Eyeriss [10] (as small as (12, 14)) and Google TPU [30] (as large as (256, 256)), respectively, simplifying the design space exploration [4], [17], [29], [52].”
BackgroundMethodYüzügüler, A. C., Sönmez, C., Drumond, M. P., Oh, Y., Falsafi, B., & Frossard, P. (2023) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3572917 · Cited by 39
“2D mesh [9, 10, 22, 51] and H-trees [33, 54] are popularly used in many DNN accelerators thanks to their relatively low hardware cost.”
BackgroundNNest: Early-Stage Design Space Exploration Tool for Neural Network Inference Accelerators
Ke, L., He, X. J., & Zhang, X. (2018) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1145/3218603.3218647 · Cited by 40
DNNExplorer: A Framework for Modeling and Exploring a Novel Paradigm of FPGA-based DNN AcceleratorPreprint
Zhang, X., Ye, H., Wang, J., Lin, Y., Xiong, J., Hwu, W., & Chen, D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.12745 · Cited by 29
MEISSA: Multiplying Matrices Efficiently in a Scalable Systolic Architecture
Asgari, B., Hadidi, R., & Kim, H. (2020) · IEEE International Conference on Computer Design: VLSI in Computers and Processors (ICCD) · doi:10.1109/iccd50377.2020.00036 · Cited by 30
A Survey on Hardware Accelerator Design of Deep Learning for Edge Devices
Samanta, A., Hatai, I., & Mal, A. K. (2024) · Wireless Personal Communications · doi:10.1007/s11277-024-11443-2 · Cited by 39
“8, Eyeriss [51, 52] proposes an energy-efficient special architecture that classifies existing NN dataflows using four taxonomies.”
InfluentialChen, C., Ding, H., Peng, H., Zhu, H., Ma, R., Zhang, P., Yan, X., Wang, Y., Wang, M., Min, H., et al. (2017) · European Solid-State Circuits Conference (ESSCIRC) · doi:10.1109/esscirc.2017.8094575 · Cited by 27
“Most of the previous works are directed to convolutional neural networks (CNNs) [1]-[3], widely successful for image processing, but few to recurrent neural networks (RNNs) [4].”
BackgroundDPS: Dynamic Precision Scaling for Stochastic Computing-based Deep Neural Networks
Sim, H., Kenzhegulov, S., & Lee, J. (2018) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3195970.3196028 · Cited by 43
PiTree: Practical Implementation of ABR Algorithms Using Decision Trees
Meng, Z., Chen, J., Guo, Y., Sun, C., Hu, H., & Xu, M. (2019) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3343031.3350866 · Cited by 37
“Proposed methods include introducing new acceleration devices [23, 42, 64], or compressing neural networks [21, 22, 43].”
BackgroundMethodDeterministic conversion rule for CNNs to efficient spiking convolutional neural networks
Yang, X., Zhang, Z., Zhu, W., Yu, S., Liu, L., & Wu, N. (2020) · Science China Information Sciences · doi:10.1007/s11432-019-1468-0 · Cited by 28
MViD: Sparse Matrix-Vector Multiplication in Mobile DRAM for Accelerating Recurrent Neural Networks
Kim, B., Chung, J., Lee, E., Jung, W., Lee, S., Choi, J., Park, J., Wi, M., Lee, S., & Ahn, J. H. (2020) · IEEE Transactions on Computers · doi:10.1109/tc.2020.2984496 · Cited by 40
A NoC-based simulator for design and evaluation of deep neural networks
Chen, K.-C. J., Ebrahimi, M., Wang, T., Yang, Y.-C., & Liao, Y.-H. (2020) · Microprocessors and Microsystems · doi:10.1016/j.micpro.2020.103145 · Cited by 35
Fast-BCNN: Massive Neuron Skipping in Bayesian Convolutional Neural Networks
Wan, Q., & Fu, X. (2020) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1109/micro50266.2020.00030 · Cited by 31
Huffman Coding Based Encoding Techniques for Fast Distributed Deep Learning
Gajjala, R. R., Banchhor, S., Abdelmoniem, A. M., Dutta, A., Canini, M., & Kalnis, P. (2020) · doi:10.1145/3426745.3431334 · Cited by 47
“Similar to RLH, by using a run-length compression (RLC), Chen et al. in [7] obtained a 1 .”
MethodArchitecture of neural processing unit for deep neural networks
Lee, K. J. (2020) · Advances in computers · doi:10.1016/bs.adcom.2020.11.001 · Cited by 43
S-FLASH: A NAND Flash-based Deep Neural Network Accelerator Exploiting Bit-level Sparsity
Kang, M., Kim, H., Shin, H., Sim, J., Kim, K., & Kim, L. (2021) · IEEE Transactions on Computers · doi:10.1109/tc.2021.3082003 · Cited by 43
SOT-MRAM Digital PIM Architecture With Extended Parallelism in Matrix Multiplication
Kim, T., Jang, Y., Kang, M., Park, B., Lee, K., & Park, J. (2022) · IEEE Transactions on Computers · doi:10.1109/tc.2022.3155277 · Cited by 28
“of weights and inputs [8], copying parameters in multiple memory rows incurs a large memory area overhead.”
BackgroundGuo, P., Zhou, N., Hou, W., & Guo, L. (2022) · Optics Express · doi:10.1364/oe.468456 · Cited by 30
Sehgal, R., Thareja, T., Xie, S., Ni, C., & Kulkarni, J. P. (2023) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2023.3235210 · Cited by 28
Precise deep neural network computation on imprecise low-power analog hardwarePreprint
Binas, J., Neil, D., Indiveri, G., Liu, S., & Pfeiffer, M. (2016) · Zurich Open Repository and Archive (University of Zurich) · doi:10.5167/uzh-132629 · Cited by 27
Boosting Convolutional Neural Networks Performance Based on FPGA Accelerator
Al-Shamma, O., Fadhel, M. A., Hameed, R. A., Alzubaidi, L., & Zhang, J. (2020) · Advances in Intelligent Systems and Computing · doi:10.1007/978-3-030-16657-1_47 · Cited by 27
Das, A., Palesi, M., Kim, J., & Pande, P. P. (2024) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2024.3445829 · Cited by 42
“Due to the consistent data flow in Deep Neural Network (DNN) operations, array-based interconnects are very common in modern DNN accelerators [125], [126], [127], [128], [129].”
BackgroundPre-Defined Sparse Neural Networks With Hardware Acceleration
Dey, S., Huang, K.-W., Beerel, P. A., & Chugg, K. M. (2019) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2019.2910864 · Cited by 37
“The vast majority of this addresses only inference given a trained model [9], [23]–[26], with few addressing hardware ar X iv :1 81 2.”
BackgroundmRNA: Enabling Efficient Mapping Space Exploration for a Reconfiguration Neural Accelerator
Zhao, Z., Kwon, H., Kuhar, S., Sheng, W., Mao, Z., & Krishna, T. (2019) · IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) · doi:10.1109/ispass.2019.00040 · Cited by 34
Self-citationCANN: Curable Approximations for High-Performance Deep Neural Network Accelerators
Hanif, M. A., Khalid, F., & Shafique, M. (2019) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3316781.3317787 · Cited by 44
“Therefore, the state-of-the-art accelerators [13][8][2][11] mainly focus on accelerating the convolutional and fully-connected layers using arrays of processing elements that can perform large number of MAC operations in parallel.”
BackgroundNguyen, D.-T., Ho, N.-M., & Chang, I. (2019) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3316781.3317915 · Cited by 27
“Although there are many custom hardware designs for DNN [3, 4], these devices are still using the same design for DRAM controller.”
MethodFast Algorithms for Quaternion-Valued Convolutional Neural Networks
Cariow, A., & Cariowa, G. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.2979682 · Cited by 30
Learning cell for superconducting neural networks
Schegolev, A. E., Klenov, N. V., Soloviev, I. I., & Tereshonok, M. V. (2020) · Superconductor Science and Technology · doi:10.1088/1361-6668/abc569 · Cited by 27
“Currently, for CMOS/MOFOS ANNs the first or second types are most often used for learning since they are much easier to implement [28, 29].”
MethodDUET: Boosting Deep Neural Network Efficiency on Dual-Module Architecture
Liu, L., Qu, Z., Deng, L., Tu, F., Li, S., Hu, X., Gu, Z., Ding, Y., & Xie, Y. (2020) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1109/micro50266.2020.00066 · Cited by 31
BitSystolic: A 26.7 TOPS/W 2b~8b NPU With Configurable Data Flows for Edge Devices
Yang, Q. X., & Li, H. H. (2020) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2020.3043778 · Cited by 27
Accelerating Inference of Convolutional Neural Networks Using In-memory Computing
Dazzi, M., Sebastian, A., Benini, L., & Eleftheriou, E. S. (2021) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2021.674154 · Cited by 32
“To this end, custom hardware platforms such as applicationspecific integrated circuits (ASICs) are being designed (Chen et al., 2016) for low power and efficient execution of DNNs.”
MethodAIDA: Associative In-Memory Deep Learning Accelerator
Garzón, E., Teman, A., Lanuzza, M., & Yavits, L. (2022) · IEEE Micro · doi:10.1109/mm.2022.3190924 · Cited by 27
STRAIT: Self-Test and Self-Recovery for AI Accelerator
Lee, H., Kim, J., Park, J., & Kang, S. (2023) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2023.3236875 · Cited by 29
“such as AI accelerators [6], [7], [8], [9], [10].”
BackgroundDuan, X., Chen, Y., Li, M., Rong, Y., Xie, R., & Han, J. S. (2023) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2023.3261060 · Cited by 29
Self-citationLin, Z., Zhang, S., Jin, Q., Xia, J., Liu, Y., Yu, K., Zheng, J., Xu, X., Fan, X., Li, K., et al. (2023) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2023.3266651 · Cited by 28
“…logic circuitry around the memory array to achieve multiple types of basic logic operations [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], such as NAND , AND , OR , and NOR , by distinguishing the discharge current or voltage formed by the simultaneous activation of multiple cells…”
BackgroundA Parallel Processing CNN Accelerator on Embedded Devices Based on Optimized MobileNet
Yang, D., & Luo, Z. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3277869 · Cited by 30
DeepTools: Compiler and Execution Runtime Extensions for RaPiD AI Accelerator
Venkataramani, S., Choi, J., Srinivasan, V., Wang, W., Zhang, J., Schaal, M., Serrano, M. J., Ishizaki, K., Inoue, H., Ogawa, E., et al. (2019) · IEEE Micro · doi:10.1109/mm.2019.2931584 · Cited by 27
Variances-constrained weighted extreme learning machine for imbalanced classification
Liu, Z., Jin, W., & Mu, Y. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.04.052 · Cited by 27
Per-Tensor Fixed-Point Quantization of the Back-Propagation AlgorithmPreprint
Sakr, C., & Shanbhag, N. R. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.11732 · Cited by 39
“DNNs’ enormous computational and parameter complexity leads to high energy consumption (Chen et al., 2017), makes their training via the stochastic gradient descent (SGD) algorithm very slow often requiring hours and days (Goyal et al., 2017), and inhibits their deployment on energy and…”
BackgroundOrdering Chaos: Memory-Aware Scheduling of Irregularly Wired Neural Networks for Edge DevicesPreprint
Ahn, B. H., Lee, J., Lin, J. M., Cheng, H.-P., Hou, J., & Esmaeilzadeh, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.02369 · Cited by 68
“We also show how SERENITY affects the off-chip memory communication, which largely affects both power and inference speed (Chen et al., 2016; Gao et al., 2017; Sharma et al., 2018).”
BackgroundInfluentialFrom Algorithm to Hardware: A Survey on Efficient and Safe Deployment of Deep Neural Networks
Geng, X., Wang, Z., Chen, C., Xu, Q., Xu, K., Jin, C., Gupta, M., Yang, X., Chen, Z., Sabry Aly, M. M., et al. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3394494 · Cited by 26
“RLC was used in [204], [217], [219], and [230].”
Agamotto: A Performance Optimization Framework for CNN Accelerator With Row Stationary Dataflow
Kim, D., Jeong, S., & Kim, J.-Y. (2023) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2023.3258411 · Cited by 33
“The throughput utilization is compared to the baseline accelerator [23] to evaluate Agamotto in finding the optimal data mapping variables.”
BackgroundMethodInfluentialYue, J., Liu, Y., Feng, X., He, Y., Wang, J., Yuan, Z., Zhan, M., Liu, J., Su, J.-W., Chung, Y.-L., et al. (2023) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2023.3324954 · Cited by 28
“16 presents the energy efficiency comparison with several digital NN chips [37], [38], [39], [40], [41], [42].”
MethodFeng, Y., Soni, A., Brightwell, G., dos Reis, M. M., Wang, Z., Wang, J., Wu, Q., & Ding, Y. (2024) · Trends in Food Science & Technology · doi:10.1016/j.tifs.2024.104555 · Cited by 29
Embedded Artificial Intelligence: A Comprehensive Literature Review
Huang, X., Wang, H., Qin, S., & Tang, S.-K. (2025) · Electronics · doi:10.3390/electronics14173468 · Cited by 38
“…specialization enables their widespread application in various EAI scenarios, including image recognition, speech processing, and natural language understanding, playing a crucial role particularly in edge computing devices where power consumption and cost are stringent requirements [62,64].”
Optimizing the data placement and transformation for multi-bank CGRA computing system
Zhao, Z., Liu, Y., Sheng, W., Krishna, T., Wang, Q., & Mao, Z. (2018) · Design, Automation, and Test in Europe · doi:10.23919/date.2018.8342173 · Cited by 27
Self-citationSynthesizable Memory Arrays Based on Logic Gates for Subthreshold Operation in IoT
Fan, X., Stuijt, J., Liu, B., & Gemmeke, T. (2018) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2018.2873026 · Cited by 26
A Resources-Efficient Configurable Accelerator for Deep Convolutional Neural Networks
Hu, X., Zeng, Y., Li, Z., Zheng, X. W., Cai, S., & Xiong, X. (2019) · IEEE Access · doi:10.1109/access.2019.2919527 · Cited by 29
Di, X., Yang, H., Jia, Y., Huang, Z., & Mao, N. (2020) · Electronics · doi:10.3390/electronics9020286 · Cited by 30
“Referring to Eyerss [22] who focused on the common convolution, the authors of [19–21] concerned removing all the invalid operations in the sparse matrix multiplication.”
BackgroundScherer, M., Rutishauser, G., Cavigelli, L., & Benini, L. (2021) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2021.3075420 · Cited by 31
Designing Efficient and High-Performance AI Accelerators With Customized STT-MRAM
Mishty, K., & Sadi, M. (2021) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2021.3105958 · Cited by 29
Pure tensor program rewriting via access patterns (representation pearl)
Smith, G. H., Liu, A., Lyubomirsky, S., Davidson, S., McMahan, J., Taylor, M. B., Ceze, L., & Tatlock, Z. (2021) · doi:10.1145/3460945.3464953 · Cited by 42
SamurAI: A Versatile IoT Node With Event-Driven Wake-Up and Embedded ML Acceleration
Miro-Panades, I., Tain, B., Christmann, J., Coriat, D., Lemaire, R., Jany, C., Martineau, B., Chaix, F., Waltener, G., Pluchart, E., et al. (2022) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2022.3198505 · Cited by 29
Q-LIC: Quantizing Learned Image Compression With Channel Splitting
Sun, H., Yu, L., & Katto, J. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3231789 · Cited by 40
An In-Memory-Computing Charge-Domain Ternary CNN Classifier
Yang, X., Zhu, K., Tang, X., Wang, M., Zhan, M., Lu, N., Kulkarni, J. P., Pan, D. Z., Liu, Y., & Sun, N. (2023) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2023.3238725 · Cited by 26
“Challenges arise from the fact that modern DNN models require millions to billions of MAC OPs, making them difficult to deploy on edge platforms.”
BackgroundInfluentialA Ternary Neural Network Computing-in-Memory Processor With 16T1C Bitcell Architecture
Jeong, H., Kim, S., Park, K., Jung, J., & Lee, K. J. (2023) · IEEE Transactions on Circuits & Systems II Express Briefs · doi:10.1109/tcsii.2023.3265064 · Cited by 26
“It is well known that CNNs involve large amounts of computations and memory footprint, limiting performance and energy efficiency with conventional von Neumann architecture due to the massive data transaction [2], [3], [4].”
BackgroundGuo, C., Li, W., Wang, J., Yu, X., Liu, X., Luvembe, A. M., Wang, C., & Jin, Q. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111580 · Cited by 28
Metadata
- Authors
- Chen, Y.-H., Krishna, T., Emer, J. S., & Sze, V.
- Affiliations
- Massachusetts Institute of Technology (US)
- Georgia Institute of Technology (US)
- Nvidia (United States) (US)
- Venue
- IEEE Journal of Solid-State Circuits
- Published
- 2017-01
- Volume
- 52
- Issue
- 1
- Pages
- 127-138
- ISSN
- 0018-9200 (ISSN-L)
- 1558-173X
- OpenAlex
- W2289252105
- Semantic Scholar
- ffdaa12ef011de9dbf43be46d45a3abcc8288965
- Cited by
- 3,247
- References
- 42
- Citations by year
- 2026: 218 · 2025: 313 · 2024: 294 · 2023: 373 · 2022: 365 · 2021: 450 · 2020: 427 · 2019: 354 · 2018: 302 · 2017: 118
- FWCI
- 182.12
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE Journal of Solid-State Circuits (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
- Keywords
- reconfigurable accelerators, deep convolutional neural networks, energy efficiency, spatial architecture, data movement reduction, data compression, processing elements, AlexNet, VGG16, convolutional layers
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.