A 0.32–128 TOPS, Scalable Multi-Chip-Module-Based Deep Neural Network Inference Accelerator With Ground-Referenced Signaling in 16 nm
The DOI matches its Crossref record; 26 of the 29 references it cites were found in a registry or an index.
Abstract
Custom accelerators improve the energy efficiency, area efficiency, and performance of deep neural network (DNN) inference. This article presents a scalable DNN accelerator consisting of 36 chips connected in a mesh network on a multi-chip-module (MCM) using ground-referenced signaling (GRS). While previous accelerators fabricated on a single monolithic chip are optimal for specific network sizes, the proposed architecture enables flexible scaling for efficient inference on a wide range of DNNs, from mobile to data center domains. Communication energy is minimized with large on-chip distributed weight storage and a hierarchical network-on-chip and network-on-package, and inference energy is minimized through extensive data reuse. The 16-nm prototype achieves 1.29-TOPS/mm2 area efficiency, 0.11 pJ/op (9.5 TOPS/W) energy efficiency, 4.01-TOPS peak performance for a one-chip system, and 127.8 peak TOPS and 1903 images/s ResNet-50 batch-1 inference for a 36-chip system.
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 [59]
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
ImageNet 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
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., & Adam, H. (2018) · arXiv · doi:10.48550/arxiv.1802.02611 · Cited by 20,063
Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., & Adam, H. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01234-2_49 · Cited by 20,054
In-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537 · cited 1×
“Duetothedeterministic structure of DNNs, fixed-function accelerators have the potential to further improve area effi-ciency, energy efficiency, and performance relative to CPUs and GPUs [6]–[12].”
BackgroundEyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks
Chen, Y.-H., Krishna, T., Emer, J. S., & Sze, V. (2017) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2016.2616357 · Cited by 3,247
Self-citationEnd to End Learning for Self-Driving Cars
Bojarski, M., Del Testa, D., Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L. D., Monfort, M., Muller, U., Zhang, J., et al. (2016) · arXiv · doi:10.48550/arxiv.1604.07316 · Cited by 3,198
EIE: efficient inference engine on compressed deep neural network
Han, S., Liu, X., Mao, H., Pu, J., Pedram, A., Horowitz, M. A., & Dally, W. J. (2016) · ACM SIGARCH Computer Architecture News · doi:10.1145/3007787.3001163 · Cited by 1,937
Self-citationMoons, B., Uytterhoeven, R., Dehaene, W., & Verhelst, M. (2017) · 2017 IEEE International Solid-State Circuits Conference (ISSCC) · doi:10.1109/isscc.2017.7870353 · Cited by 479
Research works printed but not linked: 3
- saban, 2012, Xilinx Stacked Silicon Interconnect Technology Delivers Breakthrough FPGA Capacity Bandwidth and Power Efficiency
- 2017, AMBA AXI and ACE Protocol Specification AXI3 AXI4 AXI5 ACE and ACE5
- 2019, MATCH
Cited by
Items 101–122 of 122
2020 IEEE Journal of Solid-State Circuits Best Paper Award Winner Announced [Awards]
(2022) · IEEE Solid-State Circuits Magazine · doi:10.1109/mssc.2022.3162732
2020 Index IEEE Journal of Solid-State Circuits Vol. 55
(2020) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2021.3054535
Analysis of Quantization Across DNN Accelerator Architecture Paradigms
Glint, T., Jha, C. K., Awasthi, M., & Mekie, J. (2023) · Design, Automation, and Test in Europe · doi:10.23919/date56975.2023.10136899
Efficient hardware acceleration of deep neural networks via arithmetic complexity reduction
Reggiani, E. (2023) · UPCommons institutional repository (Universitat Politècnica de Catalunya) · doi:10.5821/dissertation-2117-404668
COMPRIZE: Assessing the Fusion of Quantization and Compression on DNN Hardware Accelerators
Patel, V., Shah, N., Krishna, A., Glint, T., Ronak, A. Q., & Mekie, J. (2024) · International Conference on VLSI Design · doi:10.1109/vlsid60093.2024.00048
“If the input data is [0, 0, 0, 0, 5, 0, 0, 0, 7, 0, 0], it will be encoded as [4, 3, 2] and the original data is represented as [5, 7].”
BackgroundMethodInfluentialA 77.79 GOPs/W Retentive Network FPGA Inference Accelerator with Optimized Workload
Nian, C., Zhang, W., Farrukh, F. U. D., Niu, L., Jiang, D., Chen, F., & Zhang, C. (2024) · Annual Conference of the IEEE Industrial Electronics Society (IECON) · doi:10.1109/iecon55916.2024.10905965 · Cited by 1
APCPU: Adaptive-Pooling Compression Processing Unit for Energy-Efficient DNNs Processing
Wang, W., Zhao, W., Li, M., Li, Z., Yang, J., Huang, Z., Lin, Y., Wang, C., & Xiong, X. (2025) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas56072.2025.11044228
A Low-Noise Single-Stage Coupling Clock Multiplexer for Multi-Chip Systems
Hwang, K., Oh, K., & Lee, J.-J. (2025) · IEEE International Conference on Consumer Electronics-Asia (ICCE-Asia) · doi:10.1109/icce-asia67487.2025.11263654
“Consequently, the adoption of a multi-chip architecture is expected to become increasingly important [1], [2].”
A Dimensional Order Search Based Mapping Algorithm for Deploying DNNs Onto Many-Core Neuromorphic IC
Wang, S., Lu, R., Xiao, B., Wang, Z., Guo, Y., Sun, X., & Yu, W. (2025) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2025.3640508
“Search based algorithms, including GA [20], [29], SA [16], and greedy algorithms [5], [7], [17], struggle to identify optimal mappings within constrained search spaces as core counts increase in single neuromorphic chips.”
InfluentialA Hybrid Platform for Multi-Neurons Model with Optimum Co-Design Method
Wang, S., Qiao, K., Liang, J., Hua, N., Lu, R., Zhai, J., & Pei, J. (2025) · Big Data Mining and Analytics · doi:10.26599/bdma.2025.9020106
Zhang, Z., Guan, Y., & Ye, W. (2026) · Circuits Systems and Signal Processing · doi:10.1007/s00034-025-03490-2
Zhang, Z., Yang, X., Shen, Y., Wang, J., Shi, Y., Qu, T., Hong, Z., & Xu, J. (2026) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2026.3661990
“Several general-purpose AI accelerators have been designed to enable neural network deployment on resource-constrained platforms [7], [8], [9], using parallel computation and optimized memory architectures.”
Scope: A Scalable Merged Pipeline Framework for Multi-Chip-Module NN Accelerators
Huang, Z., Jia, H., Zou, K., & Liu, Y. (2026) · Asia and South Pacific Design Automation Conference (ASP-DAC) · doi:10.1109/asp-dac66049.2026.11420654
“92267203 and 62504161) and Deng Feng Fund.”
A Compact Energy-Efficient Unified Micro-Robotic Processor With Flexible Visual Perception
Zhang, S., Wang, Y., Chen, X., Yang, H., Liu, Y., & Jia, H. (2026) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2026.3677826
“The baseline is a heterogeneous system that integrates a representative localization accelerator [18] and a high-performance NN inference processor [43], which operate in parallel.”
HiRe: A Hierarchical Reconfigurable Architecture for Large-Scale Multichiplet DNN Accelerators
CHEN, Z., Zhang, J., Lyu, D., Chen, H., Jiang, J., Sheng, W., Zhang, C., & He, G. (2026) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2026.3676552
CHARIOT: A Communication-Aware Exploration Framework for 2.5D/3D Silicon Interposer Design
Zeng, Y., Jiao, B., Jia, S., Xiong, X., Wang, C., Tian, H., Zhu, H., Liu, Q., & Chen, C. (2026) · ACM Transactions on Design Automation of Electronic Systems · doi:10.1145/3815192
COFFEE: A Carbon-Modeling and Optimization Framework for HZO-based FeFET eNVMs
Wu, H., Chen, X., Jadhav, S., Lal, A., Pentecost, L., & Gupta, U. (2026) · Design, Automation, and Test in Europe · doi:10.23919/date69613.2026.11539619 · Cited by 3
“Recent state-of-the-art (SOTA) accelerator shows that 6T SRAM accounts for 67% of the chiplet area and contributes 56% of the operational energy consumption [6], [7].”
Communication-Aware Hybrid Parallelism Mapping for Low-Cost MCM-based DNN Accelerators
Kim, J., Park, C., Nguyen, X. T., & Lee, H. (2026) · Design, Automation, and Test in Europe · doi:10.23919/date69613.2026.11539091
“Simba [5] and Zimmer et al. [6] adopt weight-centric mapping, generating heavy C2C traffic from partial-sum exchanges.”
Morsy, M. M. (2026) · Electronics · doi:10.3390/electronics15122645 · Cited by 6
Wu, Q., Bai, Y., Fang, C., Du, Y., & Du, L. (2026) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas66217.2026.11562210
Lee, H., Burd, J., Ling, T., Yang, Y.-K., Chang, N.-S., Chen, W.-C., Chen, C., Chiueh, T., Lee, M. K.-T., Yang, C., et al. (2026) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2026.3715500
“The hardware accelerators [18], [19], [20], [21], [22] achieve excellent energy and area e ffi ciency, but they are typically optimized for fixed workloads and cannot support frequent or low-latency reconfiguration.”
Li, M., Wang, W., Li, Z., Yang, J., Huang, Z., Zhao, W., & Lin, Y. (2026) · Microelectronics Journal · doi:10.1016/j.mejo.2026.107464
Metadata
- Authors
- Zimmer, B., Venkatesan, R., Shao, Y. S., Clemons, J., Fojtik, M., Jiang, N., Keller, B., Klinefelter, A., Pinckney, N., Raina, P., Tell, S. G., Zhang, Y., Dally, W. J., Emer, J. S., Gray, C. T., Keckler, S. W., & Khailany, B.
- Affiliations
- Nvidia (United States) (US)
- University of California, Berkeley (US)
- Stanford University (US)
- ORCID
- Venue
- IEEE Journal of Solid-State Circuits
- Published
- 2020-04
- Volume
- 55
- Issue
- 4
- Pages
- 920-932
- ISSN
- 0018-9200 (ISSN-L)
- 1558-173X
- OpenAlex
- W2999356938
- Semantic Scholar
- fe261bfd9d74f97268b3a9d9b700fdf5340ef434
- Cited by
- 122
- References
- 29
- Citations by year
- 2026: 15 · 2025: 25 · 2024: 16 · 2023: 28 · 2022: 20 · 2021: 15 · 2020: 3
- FWCI
- 6.42
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Journal of Solid-State Circuits (publishedVersion)
- Funders
- U.S. Government through the DARPA CRAFT Program
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Networks and Communications › Interconnection Networks and Systems
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Keywords
- multichip modules, network-on-chip, data reuse, energy efficiency, area efficiency, 16nm CMOS
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.