Conventional Von Neumann and Neuromorphic Architecture of AI Chips
DOI:
https://doi.org/10.54097/gwgea042Keywords:
GPU, FPGA, ASIC, Brain-like Chips, Memristors, Crossbar.Abstract
In the background of the shortage of computing power in the training AI field, the current solutions and future outlooks will be shown if they tackle this dilemma. From the continuation of the traditional computer structure, Von Neumann structure, the three types of AI chips GPU, FPGA, and ASIC will be introduced and state their developments and drawbacks. Besides, a brand new solution gaining inspiration from our brain will also be discussed and introduce the fundamental electronic component to accomplish the goal. The principle of how a nerve fires will also be illustrated and based on this catch a glimpse of the neural network. Besides, The principle of memristors and with the support of crossbars, how could they be used in manufacturing AI Chips will be explained. Finally, the latest research in this field. The advantages and challenges will also be involved. Finally, a comparison of these solutions will be proposed.
Downloads
References
Zhang W , Yao P., Gao B., Liu Q., Wu D., Zhang Q., Li Y., Qin Q., Li J., Zhu Z., Cai Y., Wu D., Tang J., Qian H., Wang Y., Wu H., “Edge learning using a fully integrated neuro-inspired memristor chip,” Science 381, 6663 (2023).
Baji, T. “Evolution of the GPU Device Widely Used in AI and Massive Parallel Processing,” IEEE Electron Devices Technology and Manufacturing Conference Proceedings of Technical Papers 978,1,5386,3712 (2018).
Stephen W. K., William J. D, Brucek K, Michael G, David, “G GPUs and the future of Parallel computing,” IEEE Computer Society, 0272,1732 (2011).
Sorna M V, “AI Chips: New Semiconductor Era,” International Journal of Advanced Research in Science, Engineering and Technology 7, 8 (2020).
Li B., Gu J., Jiang W., “Artificial Intelligence (AI) chip technology review,” International Conference on Machine Learning Big Data and Business Intelligence (MLBDBI). 978,1,7281, 5094 (2019).
Samuel K. M., “Intel’s Neuromorphic System Hits 8 million Neurons, 100 Million Coming by 2020,” (2019).
Robert A. N., Richard M. V., and Sean E. S., “A Mini Review of Neuromorphic Architectures and Implementations” IEEE TRANSACTIONS ON ELECTRON DEVICES,” 63, 10 (2016).
Leon O. C., “Memristor-The Missing Circuit Element,” IEEE Transactions on circuit theory, 18, 5, (1971).
Dmitri B. S, Gregory S. S, Duncan R. S& R. S W, “The missing memristor found,” Nature, 453-1, (2008).
Xia Q, J. J. Y., “Memristive crossbar arrays for brain-inspired computing,” Nature Materials, 18, 309, 323(2019).
Mei P, F., Mei B. D., Han G. C., An K. K. L, and Ding B. S., “Training and operation of an integrated neuromorphic network based on metal-oxide memristors Letter,” Nature, 10(1038), 14441 (2018).
Yao P., Wu H., Gao B., Sukru B. E., Huang X., Zhang W., Zhang Q., Deng N., Shi L., Wong H. S. P. and Qian H., “Face classification using electronic synapses,” Nature Communication 8, 15199 (2017).
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Highlights in Science, Engineering and Technology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
How to Cite
