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 701–800 of 1,001
A Mixed Signal Architecture for Convolutional Neural Networks
Lou, Q., Pan, C., McGuinness, J., Horváth, A., Naeemi, A., Niemier, M. T., & Hu, X. S. (2019) · ACM Journal on Emerging Technologies in Computing Systems · doi:10.1145/3304110 · Cited by 28
“In the digital domain, typical approaches include using optimized dataflow to efficiently reduce the data movement overhead for the dense matrix multiplication operation [8] or implementing sparse matrix multiplication by applying pruning to the network [17].”
MethodResultInfluentialChen, X., Xu, J., & Yu, Z. (2018) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2018.2883087 · Cited by 26
“Compared to the layer by layer design [7], [8], our design obtains significant power advantage.”
BackgroundA Primer on Deep Learning Architectures and Applications in Speech Processing
Ogunfunmi, T., Ramachandran, R. P., Togneri, R. B., Zhao, Y., & Xia, X. (2019) · Circuits Systems and Signal Processing · doi:10.1007/s00034-019-01157-3 · Cited by 29
“There are acceleration solutions for GPU [15], CPU [2,52,93], FPGA and ASIC platforms [3,11,13,14,35,83,104].”
BackgroundMethodShin, D., Choi, W., Park, J., & Ghosh, S. (2019) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2019.2933862 · Cited by 30
“Recently, various energyefficient DNN accelerators have been proposed [4].”
BackgroundA Logic Compatible 4T Dual Embedded DRAM Array for In-Memory Computation of Deep Neural Networks
Yoo, T., Kim, H., Chen, Q., Kim, T. T.-H., & Kim, B. (2019) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1109/islped.2019.8824826 · Cited by 32
“While custom hardware solutions such as ASIC ML accelerators have demonstrated significantly improved energy efficiency, a large portion of the overall energy consumption is still occupied by memory access [1].”
BackgroundInfluentialImplementation of DNNs on IoT devices
Zhang, Z., & Kouzani, A. Z. (2019) · Neural Computing and Applications · doi:10.1007/s00521-019-04550-w · Cited by 33
MINT: Mixed-Precision RRAM-Based IN-Memory Training Architecture
Jiang, H., Huang, S., Peng, X., & Yu, S. (2020) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas45731.2020.9181020 · Cited by 26
Habib, G., & Qureshi, S. (2022) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2022.1004988 · Cited by 32
A survey on CNN and RNN implementations
Hoffmann, J., Navarro, O., Kästner, F., Janßen, B., & Hübner, M. (2017) · Cited by 23
Maximizing system performance by balancing computation loads in LSTM accelerators
Park, J., Kung, J., Yi, W., & Kim, J. (2018) · Design, Automation, and Test in Europe · doi:10.23919/date.2018.8341971 · Cited by 25
ZARA: A Novel Zero-free Dataflow Accelerator for Generative Adversarial Networks in 3D ReRAM
Chen, F., Song, L., Li, H. H., & Chen, Y. (2019) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3316781.3317936 · Cited by 37
“Until recently, researchers began to realize that the special mathematical operator in GAN - transposed convolution (TCONV) could not be effectively executed in traditional DNN accelerators such as Eyeriss [2] and Pipelayer [18].”
BackgroundSelf-citationDeng, C., Liao, S., & Yuan, B. (2020) · IEEE Transactions on Computers · doi:10.1109/tc.2020.2981068 · Cited by 26
An On-Chip Binary-Weight Convolution CMOS Image Sensor for Neural Networks
Kim, W.-T., Lee, H., Kim, J.-G., & Lee, B. (2020) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2020.3001838 · Cited by 29
Hsiao, S., Chen, K.-C. J., Lin, C.-C., Chang, H., & Tsai, B.-C. (2020) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2020.3015238 · Cited by 26
On the Stability of Analog ReLU Networks
Elfadel, I. M. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.3042155 · Cited by 26
Hardware design and the competency awareness of a neural network
Ding, Y., Jiang, W., Lou, Q., Liu, J., Xiong, J., Hu, X. S., Xu, X., & Shi, Y. (2020) · Nature Electronics · doi:10.1038/s41928-020-00476-7 · Cited by 24
Two Sides of the Same Coin: Boons and Banes of Machine Learning in Hardware Security
Liu, W., Chang, C.-H., Wang, X., Liu, C., Fung, J. M., Ebrahimabadi, M., Karimi, N., Meng, X., & Basu, K. (2021) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2021.3084400 · Cited by 34
Kang, S., Park, G., Kim, S., Kim, S., Han, D., & Yoo, H. (2021) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2021.3120417 · Cited by 27
STT-BSNN: An In-Memory Deep Binary Spiking Neural Network Based on STT-MRAM
Nguyen, V.-T., Trinh, Q., Zhang, R., & Nakashima, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3125685 · Cited by 27
“INTRODUCTION Deep neural networks (DNNs) in on-edge AI devices face the challenge of high energy consumption due to the requirement of a large number of tensor operations, which incurs not only a high computational workload but also large memory accesses [1]–[5].”
BackgroundLow Complexity Joint Impairment Mitigation of I/Q Modulator and PA Using Neural Networks
Wu, Y., Gustavsson, U., Amat, A. G. I., & Wymeersch, H. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126024 · Cited by 28
Row-wise Accelerator for Vision Transformer
Wang, H., & Chang, T. (2022) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas54282.2022.9869928 · Cited by 26
“However, existing deep learning hardware accelerators [7]–[9] are optimized for CNN based models, which is not suitable for vision transformer models due to the significant difference in the favored architecture structure and parameters.”
BackgroundMethodTRAM: An Open-Source Template-based Reconfigurable Architecture Modeling Framework
Qiu, Y., Cao, Y., Dai, Y., Yin, W., & Wang, L. (2022) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.1109/fpl57034.2022.00021 · Cited by 25
“Therefore, CGRA is appearing as a promising accelerator design choice in domains such as security [7], [8], multimedia [9], [10], digital signal processing [11], [12], machine learning [13], [14], and high-performance computing [15], [16].”
BackgroundStack-YOLO: A Friendly-Hardware Real-Time Object Detection Algorithm
Zheng, C. (2023) · IEEE Access · doi:10.1109/access.2023.3287101 · Cited by 30
“The study by Eyeriss [52] argues that the essential factors of MAC are filter parameters and intermediate feature mapping.”
BackgroundSemiMap: A Semi-Folded Convolution Mapping for Speed-Overhead Balance on Crossbars
Deng, L., Liang, L., Wang, G., Chang, L., Hu, X., Ma, X., Liu, L., Pei, J., Li, G., & Xie, Y. (2018) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2018.2883959 · Cited by 25
“The throughput of other two accelerators, Eyeriss [36] and DNA [42], demonstrates strong dependency on the model size (although DNA does not show the results on VGG16).”
BackgroundInfluentialStructured Weight Matrices-Based Hardware Accelerators in Deep Neural Networks: FPGAs and ASICs
Ding, C., Ren, A., Yuan, G., Ma, X., Li, J., Liu, N., Yuan, B., & Wang, Y. (2018) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/3194554.3194625 · Cited by 28
“to reduce the data dimensions and mitigate overfitting issues. Max pooling is the dominant type of pooling strategy in state-of-the-art DCNNs due to its higher overall accuracy and convergence speed [1, 2]. The majority of computations occur in CONV and FC layers, while the POOL layer has a lower computational complex…”
BackgroundBalasubramanian, M., & Shrivastava, A. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.3022015 · Cited by 26
“As a result, there is a renewed surge in the application of CGRAs for compute-intensive workloads including machine learning, embedded systems, and vision functionalities [5]–[7].”
BackgroundSpeeding up Convolutional Neural Networks By Exploiting the Sparsity of Rectifier UnitsPreprint
Shi, S., & Chu, X. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1704.07724 · Cited by 43
“This property gives oppotunities for energy saving [3, 21] and MAC reducing [1] by skipping the insignificiant calculation of zero.”
BackgroundCôté‐Allard, U., Fall, C. L., Drouin, A., Campeau‐Lecours, A., Gosselin, C., Glette, K., Laviolette, F., & Gosselin, B. J. (2018) · arXiv (Cornell University) · Cited by 25
Recent Developments in Low-Power AI Accelerators: A Survey
Åleskog, C., Grahn, H., & Borg, A. (2022) · Algorithms · doi:10.3390/a15110419 · Cited by 36
“With notable AI accelerators like ICT’s DianNao [12] and its successors (DaDianNao [13], ShiDianNao [14], PuDianNao [15]), IBM’s TrueNorth [16,17], and MIT’s Eyeriss [18] paving the way for the future of AI accelerator research.”
BackgroundM-ANT: Efficient Low-bit Group Quantization for LLMs via Mathematically Adaptive Numerical Type
Hu, W., Zhang, H., Guo, C., Yu, F., Guan, R., Hua, Z., Liu, Z., Guan, Y., Guo, M., & Leng, J. (2025) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca61900.2025.00086 · Cited by 45
“We leverage the dataflow [8], [33], [37], [49], [92], [93] of matrix multiplication to hide latency.”
Field-programmable photonic nonlinearity
Wu, T., Li, Y., Ge, L., & Feng, L. (2025) · Nature Photonics · doi:10.1038/s41566-025-01660-x · Cited by 47
“Compared with the general-purpose CPUs and GPUs, reconfigurable computing-based accelerators can execute computation-intensive operations—for example, matrix-vector multiplication or convolutions that are the fundamental operations used in artificial neural networks 3,4 —orders of magnitude faster.”
INVITED: A Modular Digital VLSI Flow for High-Productivity SoC Design
Khailany, B., Krimer, E., Venkatesan, R., Clemons, J., Emer, J., Fojtik, M. R., Klinefelter, A., Pellauer, M., Pinckney, N., Shao, Y. S., et al. (2018) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac.2018.8465897 · Cited by 24
Self-citationRomaszkan, W., Li, T., Melton, T., Pamarti, S., & Gupta, P. (2020) · Design, Automation, and Test in Europe · doi:10.23919/date48585.2020.9116289 · Cited by 30
“Unlike conventional fixedpoint binary accelerators which rely on tightly coupling compute units with small, local scratchpads for cheap intermediate value buffering [6, 25, 26], ACOUSTIC leverages much larger available compute density to eliminate partial sums and amortize SC to binary conversion costs (otherwise part…”
BackgroundMethodMARS: Multimacro Architecture SRAM CIM-Based Accelerator With Co-Designed Compressed Neural Networks
Sie, S.-H., Lee, J.-L., Chen, Y.-R., Yeh, Z.-W., Li, Z., Lu, C.-C., Hsieh, C.-C., Chang, M., & Tang, K. (2021) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2021.3082107 · Cited by 28
“Therefore, several studies have focused on how to implement fine-grained pruning efficiently on hardware because of its higher compression rate [5], [34] [35].”
BackgroundMethodSelf-citationPreVIous: A Methodology for Prediction of Visual Inference Performance on IoT Devices
Velasco Montero, D., Fernández Berni, J., Carmona Galán, R., & Rodríguez Vázquez, Á. B. (2020) · LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) · Cited by 24
XOR-CIM: compute-in-memory SRAM architecture with embedded XOR encryption
Huang, S., Jiang, H., Peng, X., Li, W., & Yu, S. (2020) · doi:10.1145/3400302.3415678 · Cited by 42
“TPU [12] or Eyeriss [13]).”
The Why, What, and How of Artificial General Intelligence Chip Development
James, A. P. (2021) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2021.3069871 · Cited by 31
“Some of the implementations are Diannao [53], Dadiannao [54], Pudi-4 annao [55], Shidiannao [56], Eyeriss [57], EIE [58], Origami [59], Envision [60], TPU [61], Tesla, DPU, Q4MobilEye, Parker, S32V234, and Myriad 2 [62].”
Batch Normalization Processor Design for Convolution Neural Network Training and Inference
Ting, Y.-S., Teng, Y.-F., & Chiueh, T. (2021) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas51556.2021.9401434 · Cited by 28
Cohmeleon: Learning-Based Orchestration of Accelerator Coherence in Heterogeneous SoCs
Zuckerman, J., Giri, D., Kwon, J., Mantovani, P., & Carloni, L. P. (2021) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1145/3466752.3480065 · Cited by 28
“A multitude of accelerators have been designed for many different application domains [12, 17, 31, 39, 47, 60, 62, 65, 66, 76, 81, 86, 90].”
BackgroundSMIV: A 16-nm 25-mm² SoC for IoT With Arm Cortex-A53, eFPGA, and Coherent Accelerators
Lee, S. K., Whatmough, P. N., Donato, M., Ko, G. G., Brooks, D. M., & Wei, G.-Y. (2021) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2021.3115466 · Cited by 30
Algorithm-Hardware Co-Design for Efficient Brain-Inspired Hyperdimensional Learning on Edge
Ni, Y., Kim, Y., Rosing, T., & Imani, M. (2022) · Design, Automation, and Test in Europe · doi:10.23919/date54114.2022.9774524 · Cited by 32
An Eight-Core RISC-V Processor With Compute Near Last Level Cache in Intel 4 CMOS
Chen, G. K., Knag, P. C., Tokunaga, C., & Krishnamurthy, R. K. (2022) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2022.3228765 · Cited by 29
Dependable DNN Accelerator for Safety-Critical Systems: A Review on the Aging Perspective
Moghaddasi, I., Gorgin, S., & Lee, J. (2023) · IEEE Access · doi:10.1109/access.2023.3300376 · Cited by 30
Yi, W., Mo, K., Wang, W., Zhou, Y.-T., Zeng, Y., Yuan, Z., Cheng, B., & Pan, B. (2024) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2024.3350664 · Cited by 27
“At the falling edge of the second clock cycle, cell A is chosen with RWL [ 7 ] set to 1.”
BackgroundMethodHardware for Deep Learning Acceleration
Song, C., Ye, C., Sim, Y., & Jeong, D. S. (2024) · Advanced Intelligent Systems · doi:10.1002/aisy.202300762 · Cited by 29
Agrawal, A., Ankit, A., & Roy, K. (2018) · IEEE Transactions on Computers · doi:10.1109/tc.2018.2867048 · Cited by 26
“To overcome this bottleneck, there has been intense research for reducing data movements [7].”
BackgroundGeneSys: Enabling Continuous Learning through Neural Network Evolution in Hardware
Samajdar, A., Mannan, P., Garg, K., & Krishna, T. (2018) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1109/micro.2018.00074 · Cited by 28
“Hardware acceleration of neural networks is a hot research topic with a lot of architecture choices [17], [49]–[56] and silicon implementations [25]–[28].”
BackgroundSelf-citationInterpolation-Based Object Detection Using Motion Vectors for Embedded Real-time Tracking Systems
Ujiie, T., Hiromoto, M., & Sato, T. G. (2018) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2018.00104 · Cited by 26
“Realtime detection hence has been significantly difficult, but real-time implementations of CNNs are recently realized by the development of efficient models [17, 23, 24] and the application of highly parallelized computational resources, such as GPUs and specialized hardwares [3, 11, 21].”
MethodCompute-Efficient Neural-Network Acceleration
Wu, E., Zhang, X., Berman, D., Cho, I., & Thendean, J. (2019) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/3289602.3293925 · Cited by 33
“In contrast to the spatial lane mapping in [3], our convolution schedule maps spatial axes to compute cycles and maps channels to lanes.”
BackgroundTowards the Development of Analog Neuromorphic Chip Prototype with 2.4M Integrated Memristors
Kataeva, I., Ohtsuka, S., Nili, H., Kim, H., Isobe, Y., Yako, K., & Strukov, D. B. (2019) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2019.8702125 · Cited by 28
Prediction of Gas Concentration Using Gated Recurrent Neural Networks
Wang, S., Hu, Y., Burgués, J., Marco, S., & Liu, S. (2020) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas48895.2020.9073806 · Cited by 24
Agrawal, A., Kosta, A. K., Kodge, S., Kim, D. E., & Roy, K. (2020) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2020.3014250 · Cited by 26
“While several digital ASICs have been explored for accelerating DNNs [13]–[15], they lack flexibility and cannot run general-purpose workloads.”
BackgroundA Systolic Dataflow Based Accelerator for CNNs
Das, S., Roy, A., Chandrasekharan, K. K., Deshwal, A., & Lee, S. (2020) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas45731.2020.9180403 · Cited by 25
Wang, Z., Kara, K., Zhang, H., Alonso, G., Mutlu, O., & Zhang, C. (2019) · Proceedings of the VLDB Endowment · doi:10.14778/3317315.3317322 · Cited by 30
“Other research [2,4,7,10,11,12,16,20,29, 30,38,39,41,41,54,62,77,105] focuses on using a fixed-point, lowprecision data representation and arithmetic to accelerate DNNs, instead of using full-precision.”
BackgroundNeuromorphic Hardware Accelerator for SNN Inference based on STT-RAM Crossbar Arrays
Kulkarni, S. R., Kadetotad, D. V., Yin, S., Seo, J.-S., & Rajendran, B. (2019) · 2019 26th IEEE International Conference on Electronics, Circuits and Systems (ICECS) · doi:10.1109/icecs46596.2019.8964886 · Cited by 23
“The growing demand for on-device intelligence in mobile and edge devices has necessitated the efforts to design alternative learning paradigms and hardware accelerators [4], [5].”
BackgroundYin, G., Chen, Y., Zhou, M., Tang, W., Lee, M., Yang, Z., Liao, T., Du, X., Narayanan, V., Yang, H., et al. (2023) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2023.3326955 · Cited by 23
“In response to the escalating data and computing demands, many accelerators have been proposed, with a particular interest in the acceleration of multiply-accumulate (MAC) operations [2], [3], [4], [5], [6].”
BackgroundSelf-citationWu, L., Hwang, B., & Jiang, J. (2025) · Chemical Engineering Journal · doi:10.1016/j.cej.2025.163801 · Cited by 25
SCALE-Sim V3: a Modular Cycle-Accurate Systolic Accelerator Simulator for End-To-End System Analysis
Raj, R., Banerjee, S., Chandra, N., Wan, Z., Tong, J., Samajdhar, A., & Krishna, T. (2025) · IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) · doi:10.1109/ispass64960.2025.00026 · Cited by 25
“In addition, it supports flexibility in mapping operators to the systolic array via three classic dataflows [27] - input stationary, weight stationary, and output stationary.”
Self-citationParana: A Parallel Neural Architecture Considering Thermal Problem of 3D Stacked Memory
Yin, S., Tang, S., Lin, X., Ouyang, P., Tu, F., Liu, L., Zhao, J., Xu, C., Li, S., Xie, Y., et al. (2018) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2018.2858230 · Cited by 24
“3 reviews the spatial architecture in ShiDianNao [29], Eyeriss [8], and ENVISION [50].”
BackgroundMethodInfluentialBenchmark of Ferroelectric Transistor-Based Hybrid Precision Synapse for Neural Network Accelerator
Luo, Y., Wang, P., Peng, X., Sun, X., & Yu, S. (2019) · IEEE Journal on Exploratory Solid-State Computational Devices and Circuits · doi:10.1109/jxcdc.2019.2925061 · Cited by 25
Ghodrati, S., Sharma, H., Kinzer, S., Yazdanbakhsh, A., Park, J., Kim, N. S., Burger, D., & Esmaeilzadeh, H. (2020) · International Conference on Parallel Architectures and Compilation Techniques (PACT) · doi:10.1145/3410463.3414634 · Cited by 29
Sun, B., Yang, L., Dong, P., Zhang, W., Dong, J., & Young, C. (2018) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2018.00219 · Cited by 26
An Efficient Accelerator for Multiple Convolutions From the Sparsity Perspective
Chen, Q., Huang, Y., Sun, R., Song, W., Lu, Z., Fu, Y., & Li, L. (2020) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2020.2976454 · Cited by 22
ESSA: An energy-Aware bit-Serial streaming deep convolutional neural network accelerator
Hsu, L.-C., Chiu, C., Lin, K.-T., Chou, H.-H., & Pu, Y.-Y. (2020) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2020.101831 · Cited by 24
Efficient Hardware Architectures for 1D- and MD-LSTM Networks
Rybalkin, V., Sudarshan, C., Weis, C., Lappas, J., Wehn, N., & Cheng, L. (2020) · Journal of Signal Processing Systems · doi:10.1007/s11265-020-01554-x · Cited by 25
Kim, M., & Seo, J.-S. (2020) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2020.3029235 · Cited by 25
“DRAM access comparison to (a) [13] and [15] at each convolution layer and (b) [13] and [15], [16] in the total convolution layers for VGG-16.”
BackgroundMethodResultInfluentialFlash Memory Array for Efficient Implementation of Deep Neural Networks
Han, R., Xiang, Y., Huang, P. C., Shan, Y., Liu, X., & Kang, J. (2020) · Advanced Intelligent Systems · doi:10.1002/aisy.202000161 · Cited by 23
Xia, Z., Chen, J., Huang, Q., Luo, J., & Hu, J. (2020) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2020.3039346 · Cited by 27
Hardware accelerator systems for artificial intelligence and machine learning
(2020) · Advances in computers · doi:10.1016/bs.adcom.2020.11.005 · Cited by 24
S2Engine: A Novel Systolic Architecture for Sparse Convolutional Neural Networks
Yang, J., Fu, W., Cheng, X., Ye, X., Dai, P., & Zhao, W. S. (2021) · IEEE Transactions on Computers · doi:10.1109/tc.2021.3087946 · Cited by 27
LOMA: Fast Auto-Scheduling on DNN Accelerators through Loop-Order-based Memory Allocation
Symons, A., Mei, L., & Verhelst, M. (2021) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas51828.2021.9458493 · Cited by 38
“Conventional architectures such as Eyeriss [11] unroll the scratchpads into each PE, with dedicated I, W and O links.”
MethodEnabling on-device classification of ECG with compressed learning for health IoT
Li, W., Chu, H., Huang, B., Huan, Y., Zheng, L., & Zou, Z. (2021) · Microelectronics Journal · doi:10.1016/j.mejo.2021.105188 · Cited by 23
C-Testing and Efficient Fault Localization for AI Accelerators
Chaudhuri, A., Liu, C., Fan, X., & Chakrabarty, K. (2021) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2021.3107401 · Cited by 23
ApproxNet: Content and Contention-Aware Video Object Classification System for Embedded Clients
Xu, R., Kumar, R., Wang, P., Bai, P., Meghanath, G., Chaterji, S., Mitra, S., & Bagchi, S. (2021) · ACM Transactions on Sensor Networks · doi:10.1145/3463530 · Cited by 43
“MCDNN [18], being one of most representative works, and well-known DNNs like ResNet, MobileNets [22, 23] all fall into this category.”
BackgroundInfluentialRED++ : Data-Free Pruning of Deep Neural Networks via Input Splitting and Output Merging
Yvinec, E., Dapogny, A., Cord, M., & Bailly, K. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3179616 · Cited by 38
IECA: An In-Execution Configuration CNN Accelerator With 30.55 GOPS/mm² Area Efficiency
Huang, B., Huan, Y., Chu, H., Xu, J., Liu, L., Zheng, L., & Zou, Z. (2021) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2021.3108762 · Cited by 27
Efficient Fault-Criticality Analysis for AI Accelerators using a Neural Twin
Chaudhuri, A., Chen, C.-Y., Talukdar, J., Madala, S., Dubey, A. K., & Chakrabarty, K. (2021) · IEEE International Test Conference (ITC) · doi:10.1109/itc50571.2021.00015 · Cited by 24
Lee, S. S., Nguyen, T. D., Meher, P. K., & Park, S. Y. (2022) · IEEE Access · doi:10.1109/access.2022.3162066 · Cited by 28
Lall, A., & Tallur, S. G. (2023) · Scientific Reports · doi:10.1038/s41598-023-29277-6 · Cited by 26
“To improve the performance, we implemented a CNN-layer accelerator to perform the convolution operation based on the row-stationary approach, inspired by the Eyeriss architecture 35 , which is an energy-efficient reconfigurable accelerator for deep convolutional neural networks.”
MethodThe Hardware Impact of Quantization and Pruning for Weights in Spiking Neural Networks
Schaefer, C. J., Taheri, P., Horeni, M., & Joshi, S. (2023) · IEEE Transactions on Circuits & Systems II Express Briefs · doi:10.1109/tcsii.2023.3260701 · Cited by 32
“We use an Eyeriss-Like architecture [16] illustrated in Figure 2.”
MethodAn Efficient Accelerator Based on Lightweight Deformable 3D-CNN for Video Super-Resolution
Zhang, S., Mao, W., & Wang, Z. (2023) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2023.3258446 · Cited by 27
“In terms of the computing system, we employ a row-stationary dataflow [39] to implement channel-wise processing for DwConv2D and process Pox = 4 rows in parallel.”
MethodAn Empirical Approach to Enhance Performance for Scalable CORDIC-Based Deep Neural Networks
Raut, G., Karkun, S., & Vishvakarma, S. K. (2023) · ACM Transactions on Reconfigurable Technology and Systems · doi:10.1145/3596220 · Cited by 32
“The state-of-the-art techniques for efficient hardware design have been presented in [6, 28, 43].”
MethodNonlinear optical feature generator for machine learning
Yildirim, M., Oğuz, İ., Kaufmann, F., Escalé, M. R., Grange, R., Psaltis, D., & Moser, C. (2023) · APL Photonics · doi:10.1063/5.0158611 · Cited by 39
Reza, M. F., Le, T., De, B., Bayoumi, M., & Zhao, D. (2018) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2018.8351580 · Cited by 28
True Gradient-Based Training of Deep Binary Activated Neural Networks Via Continuous Binarization
Sakr, C., Choi, J., Wang, Z., Gopalakrishnan, K., & Shanbhag, N. R. (2018) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2018.8461456 · Cited by 35
“Taking advantage of the sparsity also enables efficient implementation via zero-skipping [3].”
BackgroundA High-Accuracy Implementation for Softmax Layer in Deep Neural Networks
Alabassy, B., Safar, M., & El-Kharashi, M. W. (2020) · International Conference on Design & Technology of Integrated Systems in Nanoscale Era (DTIS) · doi:10.1109/dtis48698.2020.9081313 · Cited by 25
“Many hardware accelerator designs were proposed on different platforms from clusters of Graphics Processing Units (GPUs) [4] to more energy-efficient solutions using Field-Programmable Gate Arrays (FPGAs) [5, 6] and Application Specific Integrated Circuits (ASICs) [7, 8].”
BackgroundSSM: a high-performance scheme for in situ training of imprecise memristor neural networks
Wang, Y., Wu, S., Tian, L., & Shi, L. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.04.130 · Cited by 28
Reducing Energy Bloat in Large Model Training
Chung, J.-W., Gu, Y., Jang, I., Meng, L., Bansal, N., & Chowdhury, M. (2024) · ACM SIGOPS Symposium on Operating Systems Principles (SOSP) · doi:10.1145/3694715.3695970 · Cited by 65
“While energy optimization is well-studied in the hardware community [7, 14, 75, 84], the power bottleneck of recent datacenters [9, 11, 12, 53, 65] shows that efficiency gains from hardware advancement alone are not sufficient to sustain the growing demand for AI compute.”
WSC-LLM: Efficient LLM Service and Architecture Co-exploration for Wafer-scale Chips
Xu, Z. M., Kong, D., Liu, J., Li, J., Hou, J., Dai, X., Li, C., Wei, S., Hu, Y., & Yin, S. (2025) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1145/3695053.3731101 · Cited by 43
“The de-sign of the PE array and its dataflows are derived from existing works [14, 17, 19, 27, 41, 43, 74, 80, 90].”
An energy-efficient and high-throughput bitwise CNN on sneak-path-free digital ReRAM crossbar
Ni, L., Liu, Z., Song, W., Yang, J. J., Yu, H., Wang, K., & Wang, Y. (2017) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1109/islped.2017.8009177 · Cited by 25
Neural networks: Efficient implementations and applications
Zhang, C., & Xu, W. (2017) · IEEE International Conference on ASIC (ASICON) · doi:10.1109/asicon.2017.8252654 · Cited by 22
“Eyeriss [26] is the other systolicbased accelerator for CNNs.”
BackgroundEfficient fast convolution architectures for convolutional neural network
Xu, W., Wang, Z., You, X., & Zhang, C. (2017) · IEEE International Conference on ASIC (ASICON) · doi:10.1109/asicon.2017.8252623 · Cited by 23
“Paper [7] focused on the local data reuse of filters and feature map pixels, which maintains high level parallelism and minimizes data movement inside the network.”
Background(2018) · Cited by 22
DNN pruning and mapping on NoC-Based communication infrastructure
Mirmahaleh, S. Y. H., & Rahmani, A. M. (2019) · Microelectronics Journal · doi:10.1016/j.mejo.2019.104655 · Cited by 27
Li, T., Ma, Y., & Endoh, T. (2020) · IEEE Access · doi:10.1109/access.2020.3013934 · Cited by 22
Roofline-Model-Based Design Space Exploration for Dataflow Techniques of CNN Accelerators
Park, C., Park, S., & Park, C. S. (2020) · IEEE Access · doi:10.1109/access.2020.3025550 · Cited by 25
GNNerator: A Hardware/Software Framework for Accelerating Graph Neural Networks
Stevens, J. R., Das, D., Avancha, S., Kaul, B., & Raghunathan, A. (2021) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac18074.2021.9586122 · Cited by 24
A Review of FPGA‐Based Custom Computing Architecture for Convolutional Neural Network Inference
Xiyuan, P., Yu, J., Yao, B., Liansheng, L., & Peng, Y. (2021) · Chinese Journal of Electronics · doi:10.1049/cje.2020.11.002 · Cited by 24
HeSA: Heterogeneous Systolic Array Architecture for Compact CNNs Hardware Accelerators
Xu, R., Ma, S., Wang, Y., & Guo, Y. (2021) · Design, Automation, and Test in Europe · doi:10.23919/date51398.2021.9474145 · Cited by 23
FPGA-based Acceleration for Convolutional Neural Networks on PYNQ-Z2
Huynh, T. V. (2022) · International Journal of Computing and Digital Systems · doi:10.12785/ijcds/110136 · Cited by 25
“With respect to hardware implementations, three hardware platforms can be used as CNN accelerators: graphic processing units (GPU) [5], application-specific integrated circuits (ASICs) [6], and field programmable gate arrays (FPGAs) [7-8].”
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.