Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

A carbon-nanotube-based tensor processing unit

Nature Electronics volume 7, pages 684–693 (2024)Cite this article

Subjects

Abstract

The growth of data-intensive computing tasks requires processing units with higher performance and energy efficiency, but these requirements are increasingly difficult to achieve with conventional semiconductor technology. One potential solution is to combine developments in devices with innovations in system architecture. Here we report a tensor processing unit (TPU) that is based on 3,000 carbon nanotube field-effect transistors and can perform energy-efficient convolution operations and matrix multiplication. The TPU is constructed with a systolic array architecture that allows parallel 2 bit integer multiply–accumulate operations. A five-layer convolutional neural network based on the TPU can perform MNIST image recognition with an accuracy of up to 88% for a power consumption of 295 µW. We use an optimized nanotube fabrication process that offers a semiconductor purity of 99.9999% and ultraclean surfaces, leading to transistors with high on-current densities and uniformity. Using system-level simulations, we estimate that an 8 bit TPU made with nanotube transistors at a 180 nm technology node could reach a main frequency of 850 MHz and an energy efficiency of 1 tera-operations per second per watt.

Access through your institution
Buy or subscribe

This is a preview of subscription content, access via your institution

Access options

Access through your institution

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: CNT FET-based digit computing system for tensor processing acceleration.
Fig. 2: Electrical characteristics of a top-gated p-FET and basic logic gates.
Fig. 3: PE and data flow for convolution in the CNT TPU.
Fig. 4: Image edge extraction with single and combined kernels.
Fig. 5: Five-layer CNN with the CNT TPU, performance metrics and comparison of different systems.
Fig. 6: Matrix multiplication by the CNT TPU.

Similar content being viewed by others

Data availability

The data that support the plots within this paper and the other findings of this study are available from the corresponding author upon reasonable request.

Code availability

The custom codes for this study are available from the corresponding author upon request.

References

  1. Ionescu, A. M. Energy efficient computing and sensing in the zettabyte era: from silicon to the cloud. In Proc. 2017 IEEE International Electron Devices Meeting (IEDM) 1.2.1–1.2.8 (IEEE, 2017); https://doi.org/10.1109/IEDM.2017.8268307

  2. Li, H., Ota, K. & Dong, M. Learning IoT in edge: deep learning for the Internet of Things with edge computing. IEEE Netw. 32, 96–101 (2018).

    Article  Google Scholar 

  3. Service, R. F. Is silicon’s reign nearing its end? Science 323, 1000–1002 (2009).

    Article  Google Scholar 

  4. Markov, I. L. Limits on fundamental limits to computation. Nature 512, 147–154 (2014).

    Article  Google Scholar 

  5. Dean, J., Patterson, D. & Young, C. A new golden age in computer architecture: empowering the machine-learning revolution. IEEE Micro 38, 21–29 (2018).

    Article  Google Scholar 

  6. Qiu, C. et al. Scaling carbon nanotube complementary transistors to 5-nm gate lengths. Science 355, 271–276 (2017).

    Article  Google Scholar 

  7. Liu, L. et al. Aligned, high-density semiconducting carbon nanotube arrays for high-performance electronics. Science 368, 850–856 (2020).

    Article  Google Scholar 

  8. Hills, G. et al. Modern microprocessor built from complementary carbon nanotube transistors. Nature 572, 595–602 (2019).

    Article  Google Scholar 

  9. Franklin, A. D. et al. Sub-10 nm carbon nanotube transistor. Nano Lett. 12, 758–762 (2012).

    Article  Google Scholar 

  10. Sabry Aly, M. M. et al. The N3XT approach to energy-efficient abundant-data computing. Proc. IEEE 107, 19–48 (2019).

    Article  Google Scholar 

  11. Shulaker, M. M. et al. Three-dimensional integration of nanotechnologies for computing and data storage on a single chip. Nature 547, 74–78 (2017).

    Article  Google Scholar 

  12. Gomez-Luna, J. et al. Benchmarking memory-centric computing systems: analysis of real processing-in-memory hardware. In Proc. 12th International Green and Sustainable Computing Conference (IGSC) 1–7 (IEEE, 2021); https://doi.org/10.1109/IGSC54211.2021.9651614

  13. Mutlu, O., Ghose, S., Gómez-Luna, J. & Ausavarungnirun, R. Processing data where it makes sense: enabling in-memory computation. Microprocess. Microsyst. 67, 28–41 (2019).

    Article  Google Scholar 

  14. Kang, M., Keel, M.-S., Shanbhag, N. R., Eilert, S. & Curewitz, K. An energy-efficient VLSI architecture for pattern recognition via deep embedding of computation in SRAM. In Proc. 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 8326–8330 (IEEE, 2014); https://doi.org/10.1109/ICASSP.2014.6855225

  15. Fujiki, D., Mahlke, S. & Das, R. Duality cache for data parallel acceleration. In Proc. ACM/IEEE 46th Annual International Symposium on Computer Architecture (ISCA) 1–14 (IEEE, 2019).

  16. Seshadri, V. et al. Ambit: in-memory accelerator for bulk bitwise operations using commodity DRAM technology. In Proc. 50th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO) 273–287 (ACM, 2017).

  17. Seshadri, V. et al. RowClone: accelerating data movement and initialization using DRAM. Preprint at https://doi.org/10.48550/arXiv.1805.03502 (2018).

  18. Gómez-Luna, J. et al. Benchmarking a new paradigm: experimental analysis and characterization of a real processing-in-memory system. IEEE Access 10, 52565–52608 (2022).

    Article  Google Scholar 

  19. memBrainTM Products. Silicon Storage Technology https://www.sst.com/membraintm-products

  20. Mahmoodi, M. R. & Strukov, D. An ultra-low energy internally analog, externally digital vector-matrix multiplier based on NOR flash memory technology. In Proc. 55th Annual Design Automation Conference 1–6 (ACM, 2018); https://doi.org/10.1145/3195970.3195989

  21. Wong, H.-S. P. & Salahuddin, S. Memory leads the way to better computing. Nat. Nanotechnol. 10, 191–194 (2015).

    Article  Google Scholar 

  22. Sebastian, A., Le Gallo, M., Khaddam-Aljameh, R. & Eleftheriou, E. Memory devices and applications for in-memory computing. Nat. Nanotechnol. 15, 529–544 (2020).

    Article  Google Scholar 

  23. Yao, P. et al. Fully hardware-implemented memristor convolutional neural network. Nature 577, 641–646 (2020).

    Article  Google Scholar 

  24. Cai, F. et al. A fully integrated reprogrammable memristor–CMOS system for efficient multiply–accumulate operations. Nat. Electron. 2, 290–299 (2019).

    Article  Google Scholar 

  25. Lin, P. et al. Three-dimensional memristor circuits as complex neural networks. Nat. Electron. 3, 225–232 (2020).

    Article  Google Scholar 

  26. Zidan, M. A. et al. A general memristor-based partial differential equation solver. Nat. Electron. 1, 411–420 (2018).

    Article  Google Scholar 

  27. Hung, J.-M. et al. A four-megabit compute-in-memory macro with eight-bit precision based on CMOS and resistive random-access memory for AI edge devices. Nat. Electron. 4, 921–930 (2021).

    Article  Google Scholar 

  28. Joshi, V. et al. Accurate deep neural network inference using computational phase-change memory. Nat. Commun. 11, 2473 (2020).

    Article  Google Scholar 

  29. Ambrogio, S. et al. Equivalent-accuracy accelerated neural-network training using analogue memory. Nature 558, 60–67 (2018).

    Article  Google Scholar 

  30. Nandakumar, S. R. et al. Experimental demonstration of supervised learning in spiking neural networks with phase-change memory synapses. Sci. Rep. 10, 8080 (2020).

    Article  Google Scholar 

  31. Sarwat, S. G., Kersting, B., Moraitis, T., Jonnalagadda, V. P. & Sebastian, A. Phase-change memtransistive synapses for mixed-plasticity neural computations. Nat. Nanotechnol. 17, 507–513 (2022).

  32. Berdan, R. et al. Low-power linear computation using nonlinear ferroelectric tunnel junction memristors. Nat. Electron. 3, 259–266 (2020).

    Article  Google Scholar 

  33. Shi, Y. et al. Neuroinspired unsupervised learning and pruning with subquantum CBRAM arrays. Nat. Commun. 9, 5312 (2018).

    Article  Google Scholar 

  34. Mennel, L. et al. Ultrafast machine vision with 2D material neural network image sensors. Nature 579, 62–66 (2020).

    Article  Google Scholar 

  35. Jung, S. et al. A crossbar array of magnetoresistive memory devices for in-memory computing. Nature 601, 211–216 (2022).

    Article  Google Scholar 

  36. Torrejon, J. et al. Neuromorphic computing with nanoscale spintronic oscillators. Nature 547, 428–431 (2017).

    Article  Google Scholar 

  37. Raina, R., Madhavan, A. & Ng, A. Y. Large-scale deep unsupervised learning using graphics processors. In Proc. 26th Annual International Conference on Machine Learning – ICML ’09 1–8 (ACM, 2009); https://doi.org/10.1145/1553374.1553486

  38. Wang, C. et al. DLAU: a scalable deep learning accelerator unit on FPGA. IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst. 36, 513–517 (2016).

    Google Scholar 

  39. Chen, Y., Chen, T., Xu, Z., Sun, N. & Temam, O. DianNao family: energy-efficient hardware accelerators for machine learning. Commun. ACM 59, 105–112 (2016).

    Article  Google Scholar 

  40. Jouppi, N. P. et al. In-datacenter performance analysis of a tensor processing unit. In Proc. ACM/IEEE 44th Annual International Symposium on Computer Architecture (ISCA) 1–12 (IEEE, 2017).

  41. Kung, H. T. Why systolic architectures?. Computer 15, 37–46 (1982).

    Article  Google Scholar 

  42. Jouppi, N. P. et al. Ten lessons from three generations shaped Google’s tpuv4i: industrial product. In Proc. ACM/IEEE 48th Annual International Symposium on Computer Architecture (ISCA) 1–14 (IEEE, 2021).

  43. Hu, Y. H. & Kung, S.-Y. in Handbook of Signal Processing Systems (eds. Bhattacharyya, S. S. et al.) 817–849 (Springer, 2010); https://doi.org/10.1007/978-1-4419-6345-1_29

  44. Gysel, P., Motamedi, M. & Ghiasi, S. Hardware-oriented approximation of convolutional neural networks. Preprint at arxiv.org/abs/1604.03168 (2016).

  45. Khwa, W.-S. et al. A 65nm 4kb algorithm-dependent computing-in-memory SRAM unit-macro with 2.3ns and 55.8TOPS/W fully parallel product-sum operation for binary DNN edge processors. In Proc. 2018 IEEE International Solid - State Circuits Conference - (ISSCC) 496–498 (IEEE, 2018); https://doi.org/10.1109/ISSCC.2018.8310401

  46. Tang, J. et al. A reliable all-2D materials artificial synapse for high energy-efficient neuromorphic computing. Adv. Funct. Mater. 31, 2011083 (2021).

    Article  Google Scholar 

  47. Liu, C. et al. Complementary transistors based on aligned semiconducting carbon nanotube arrays. ACS Nano 16, 21482–21490 (2022).

    Article  Google Scholar 

  48. Zhang, Z. et al. Complementary carbon nanotube metal–oxide–semiconductor field-effect transistors with localized solid-state extension doping. Nat. Electron. 6, 999–1008 (2023).

    Article  Google Scholar 

  49. Zhao, C. et al. Exploring the performance limit of carbon nanotube network film field-effect transistors for digital integrated circuit applications. Adv. Funct. Mater. 29, 1808574 (2019).

  50. Lin, Y. et al. Enhancement-mode field-effect transistors and high-speed integrated circuits based on aligned carbon nanotube films. Adv. Funct. Mater. 32, 2104539 (2022).

    Article  Google Scholar 

  51. Lee, C.-S., Pop, E., Franklin, A. D., Haensch, W. & Wong, H.-S. P. A compact virtual-source model for carbon nanotube FETs in the sub-10-nm regime—Part I. Intrinsic elements. IEEE Trans. Electron. Devices 62, 3061–3069 (2015).

    Article  Google Scholar 

Download references

Acknowledgements

This work was supported by the National Key Research and Development Program of China (Project Nos. 2022YFB4401600 and 2021YFA1202904 to Z.Z.), the Natural Science Foundation of China (Project No. 62274006 to J.S. and Project Nos. 62225101 and U21A6004 to Z.Z.) and Peking Nanofab.

Author information

Author notes

  1. These authors contributed equally: Jia Si, Panpan Zhang.

Authors and Affiliations

  1. Key Laboratory for the Physics and Chemistry of Nanodevices and Center for Carbon-based Electronics, School of Electronics, Peking University, Beijing, China

    Jia Si, Chenyi Zhao, Lin Xu, Lijun Liu, Jianhua Jiang, Lian-Mao Peng & Zhiyong Zhang

  2. State Key Laboratory of Information Photonics and Optical Communications, Beijing University of Posts and Telecommunications, Beijing, China

    Panpan Zhang

  3. Hunan Institute of Advanced Sensing and Information Technology, Xiangtan University, Xiangtan, China

    Dongyi Lin & Zhiyong Zhang

  4. Beijing Institute of Carbon-based Integrated Circuits, Beijing, China

    Haitao Xu, Lian-Mao Peng & Zhiyong Zhang

Authors
  1. Jia Si
    View author publications

    Search author on:PubMed Google Scholar

  2. Panpan Zhang
    View author publications

    Search author on:PubMed Google Scholar

  3. Chenyi Zhao
    View author publications

    Search author on:PubMed Google Scholar

  4. Dongyi Lin
    View author publications

    Search author on:PubMed Google Scholar

  5. Lin Xu
    View author publications

    Search author on:PubMed Google Scholar

  6. Haitao Xu
    View author publications

    Search author on:PubMed Google Scholar

  7. Lijun Liu
    View author publications

    Search author on:PubMed Google Scholar

  8. Jianhua Jiang
    View author publications

    Search author on:PubMed Google Scholar

  9. Lian-Mao Peng
    View author publications

    Search author on:PubMed Google Scholar

  10. Zhiyong Zhang
    View author publications

    Search author on:PubMed Google Scholar

Contributions

Z.Z. and L.-M.P. proposed and supervised the project. J.S. designed the system and circuit and performed the fabrication and measurements. P.Z., D.L. and J.J. performed the simulations. J.S., C.Z., H.X., L.X. and L.L. analysed the data. J.S., Z.Z. and L.-M.P. co-wrote the manuscript. All authors discussed the results and commented on the manuscript.

Corresponding authors

Correspondence to Lian-Mao Peng or Zhiyong Zhang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Electronics thanks Franz Kreupl, Jinbo Pang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Figs. 1–9 and Tables 1–3.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Si, J., Zhang, P., Zhao, C. et al. A carbon-nanotube-based tensor processing unit. Nat Electron 7, 684–693 (2024). https://doi.org/10.1038/s41928-024-01211-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue date:

  • DOI: https://doi.org/10.1038/s41928-024-01211-2

This article is cited by

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing