Article Details

Abstract

In recent years, with the rapid development of Artificial Intelligence, the computational data or parameters have shown an exponential growth year by year, and the hardware design to support efficient computing has become a new trend. Due to its low computing power consumption, FPGA is one of the effective solutions to accelerate AI computing. AI models often contain a lot of MAC. This paper designs an experimental case based on FPGA which supports the MAC. This teaching case helps students to master the design method of MAC in neural networks, and can also inspire their thinking and improve students' ability to design FPGA AI accelerators, so as to lay a good professional and practical foundation for their further growth into learning talents.

Keywords

simulation teaching; experimental design; Multiply Accumulate

References

1. [1] Wang Zhifeng, Yang Xiaoyu, Wu Zhengyang Training of Intelligent Manufacturing Engineering Talents for Industry 4.0 [J]. Development and Evaluation of Higher Education, 2024, 40 (5): 105-118+124
2. [2] Fudan Consensus on the Construction of "New Engineering" [J]. Fudan Education Forum, 2017, 15 (2): 27-28
3. [3] WU J Y，WANG Y J，WANG P X，et al. A STT-Assisted SOT MRAM-Based In-memory Booth Multiplier for Neural Network Applications [J].IEEE Transactions on Nanotechnology，2024，23：29-34.
4. [4] HIROMITSU A，MASANORI H. B2N2: Resource Efficient Bayesian Neural Network Accelerator Using Bernoulli Sampler on FPGA [J].Integration，2023，89：1-8.
5. [5] KIM S，JANG B，LEE J，et al. A CNN Inference Accelerator on FPGA With Compression and Layer-chaining Techniques for Style Transfer Applications [J].IEEE Transactions on Circuits and Systems I: Regular Papers，2023，70（4）：1591-1604.
6. [6] ALAWAD M，LIN M J. Scalable FPGA Accelerator for Deep Convolutional Neural Networks with Stochastic Streaming [J].IEEE Transactions on Multi-Scale Computing Systems，2018，4（4）：888-899.
7. [7] Liu Qiang Construction of FPGA Innovation Classroom Guided by Core Literacy [J]. Journal of Electrical and Electronic Teaching, 2023, 45 (6): 29-32
8. [8] Li Ming, Yang Mingxin, Mu Pengcheng, etc Design and System Implementation of Communication baseband Transmission Algorithm on Domestic FPGA [J]. Experimental Technology and Management, 2024, 41 (2): 135-145
9. [9] Yu Xiuli, Wei Shimin, Bai Yuxuan Research on Practical Teaching Based on FPGA Course Group under the Background of New Engineering [J]. Experimental Technology and Management, 2020, 37 (8): 178-181
10. [10] COX D D，DEAN T. Neural Networks and Neuroscience-inspired Computer Vision [J].Current Biology，2014，24（18）：921-929.
11. [11] KRIEGESKORTE N，DOUGLAS P K. Cognitive Computational Neuroscience [J].Nature Neuroscience，2018（21）：1148-1160.
12. [12] TIWARI V，KHARE N. Hardware Implementation of Neural Network with Sigmoidal Activation Functions Using CORDIC [J].Microprocessors and Microsystems，2015，39（6）：373-381.
13. [13] PARAB J S，GAD R S，NAIK G M. Getting Hands on Altera Quartus II Software [M].Cham：Springer Nature，2018.
14. [14] ALBATOOSH A H，SHUJA'A M I，AL-NEDAWE B M. A VHDL Code for Offset Pulse Position Modulation Working with Reed Solomon System by Using ModelSim [J].Journal of Techniques，2022，4（4）：50-60.
15. [15] Chen Yan, Liu Junwen, Gu Guolong, etc Experimental Teaching Design of Addition Subtraction Multiplexing Circuit Based on Quartus II and Modelsim [J]. Modern Information Technology, 2023, 7 (13): 181-188

Funding

Higher Education Teaching Quality and Teaching Reform Project (2023jxgg20, 2020jxgg10); College Student Innovation and Entrepreneurship Training Program (202314278073)

About the authors

陈艳（1991—），女，汉族，湖南常德人，讲师，博士，研究方向：面向人工智能的高性能计算；庄冰妮（2004—），女，汉族，广东揭阳人，本科在读，研究方向：电子信息工程；樊亚妮（1975—），女，汉族，陕西延川人，讲师，硕士，研究方向：教学方法与策略；杨飒（1970—），女，汉族，湖南湘潭人，教授，硕士，研究方向：信息处理与电路设计。