Influence of parameters of crossbars with STT-MRAM on the accuracy of analog neural networks

封面

如何引用文章

全文:

开放存取 开放存取
受限制的访问 ##reader.subscriptionAccessGranted##
受限制的访问 订阅或者付费存取

详细

Matrices of STT-MRAM cells assembled into a 3×3 crossbar architecture were studied. We experimentally found how the interconnect resistance affects the fidelity of the combined multiply-accumulate operations. The obtained results were compared with the simulation results in the CAD Cadence. The results of the work allow us to evaluate the contribution of the fidelity of performing combined multiply-accumulate operations, which must be considered when training analog neural networks.

作者简介

K. Kiseleva

Skolkovo Institute of Science and Technology; New Spintronic Technologies LLC

Email: ksenita.kiseleva@skoltech.ru
Skolkovo, Russia; Skolkovo, Russia

D. Cherkasov

New Spintronic Technologies LLC

Skolkovo, Russia

G. Kichin

New Spintronic Technologies LLC

Skolkovo, Russia

V. Antonov

Skolkovo Institute of Science and Technology

Skolkovo, Russia

K. Zvezdin

New Spintronic Technologies LLC

Skolkovo, Russia

参考

  1. Hamsa S., Thangadurai. N., Ananth A.G. // IJEAT. 2019. V. 8. No. 5. P. 979.
  2. Ielmini D., Pedretti G. // Adv. Intel. Syst. 2020. V. 2. No. 7. Art. No. 2000040.
  3. Saha R., Pundir Y.P., Pal P.K. et al. // J. Magn. Magn. Mater. 2022. V. 551. Art. No. 169161.
  4. Jung S., Lee H., Myung S. et al. // Nature. 2022. V. 601. P. 211.
  5. Rzeszut P., Checinski J., Rzeszut P. et al. // Sci. Reports. 2022. V. 12. No. 1. Art. No. 7178.
  6. Apalkov D., Khalkovskiy A. // ACM J. Emerg. Tech. Comp. Syst. 2013. V. 9. P. 2.
  7. Camus V., Mei L. // IEEE JETCAS. 2019. V. 9. P. 4.
  8. Lobkova M.D., Skirdkov P.N., Kichin G.A., Zvezdin K.A. // Proc. ILCon 2024 (Saint Petersburg, 2024). P. 552.
  9. Красинов Г.Я. Конструктивно-технологические особенности субмикронных МОП-транзисторов. М.: Техносфера, 2011.

补充文件

附件文件
动作
1. JATS XML
下载

版权所有 © Russian Academy of Sciences, 2025