XNOR-Popcount, an Alternative Solution to the Accumulation Multiplication Method for Approximate Computations, to Improve Latency and Power Efficiency

Các tác giả

Email tác giả liên hệ:

khoapv@hcmute.edu.vn

DOI:

https://doi.org/10.54644/jte.2025.1537

Từ khóa:

Multiply–accumulate operation, XNOR-popcount, Adder, Latency, Power consumption

Tóm tắt

Convolutional operations on neural networks are computationally intensive tasks that require significant processing time due to their reliance on calculations from multiplication circuits. In binarized neural networks, XNOR-popcount is a hardware solution designed to replace the conventional multiplied accumulator (MAC) method, which uses complex multipliers. XNOR-popcount helps optimize design area, reduce power consumption, and increase processing speed. This study implements and evaluates the performance of the XNOR-popcount design at the transistor-level on the Cadence circuit design software using 90nm CMOS technology. Based on the simulation results, for the same computational function, if MAC operation uses XNOR-popcount, the power consumption, processing time, and design complexity can be maximally reduced by up to 69%, 50%, and 48% respectively when compared to the method using conventional multipliers. Thus, the XNOR-popcount design is a useful method to apply to edge-computing platforms with minimalist hardware design, small memory space, and limited power supply.

Tải xuống: 0

Dữ liệu tải xuống chưa có sẵn.

Tiểu sử của Tác giả

Van-Khoa Pham, Ho Chi Minh City University of Technology and Education, Vietnam

Pham Van Khoa received his B.S. and M. S. E. E. degrees in Computer Technology and Electronics Engineering from the University of Technology and Education, HCM City, Vietnam, in 2010 and 2014, respectively. In 2019, he obtained his Ph.D. in Electronics Engineering from Kookmin University (K.M.U.) in Seoul, Korea. In 2010, he joined the Integrated Circuit Design Research and Education Center (I.C.D.R.E.C.), contributing to developing VN8-01 MCU, the first commercially designed and fabricated microcontroller in Vietnam. From May 2011 to 2021, he was a member of the Faculty of Electrical and Electronics Engineering at Technology and Education, HCM City, Vietnam (H.C.M.U.T.E.), and currently holds the position of senior lecturer in the Department of Computer and Communication Engineering. Presently, he serves as the Head of Computer Technology Engineering at the Faculty of International Education, H.C.M.U.T.E. His research interests encompass low-power VLSI, memory design, Internet-of-Things (IoT) and power I.C. design. He has published research papers in a variety of prestigious journals, conferences, such as, Electronics Letters, IEEE Transactions on Nanotechnology, Journal of Semiconductor Technology and Science, Micromachines, International Journal of Computing, Indonesian Journal of Electrical Engineering and Computer Science, IEEE International Symposium on Circuits and Systems (ISCAS). Email: khoapv@hcmute.edu.vn. ORCID:  https://orcid.org/0000-0002-6129-5856

Lai Le, Renesas Design Vietnam

Le Lai received his B.S. degrees in Computer Technology from the University of Technology and Education, Hochiminh City, Vietnam, in 2023. His research interests power I.C. design. Phone: 0585-856-578. Email: laile.science@gmail.com. ORCID:  https://orcid.org/0009-0002-9434-0374

Thanh-Kieu Tran Thi, Ho Chi Minh City University of Technology and Education, Vietnam

Tran Thi Thanh Kieu obtained her Master’s degree in TESOL from the University of Southern Queensland in 2015. She has been a full-time lecturer at the Faculty of Foreign Languages at Ho Chi Minh City University of Technology and Education for over twelve years. During her time there, she has taught a variety of English courses for both majors and non-majors, including general English, English for Academic Purposes (EAP), with a focus on reading and writing skills, and exam preparation. Her research interests includes English linguistics, academic writing, and professional development for teachers. Email: kieuttt@hcmute.edu.vn. ORCID:  https://orcid.org/0009-0004-7561-6122

Tài liệu tham khảo

E. Nurvitadhi et al., "Can FPGAs beat GPUs in accelerating next-generation deep neural networks?" in Proceedings of the 2017 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays, pp. 5-14. DOI: https://doi.org/10.1145/3020078.3021740

J. Chen and X. Ran. "Deep learning with edge computing: A review," Proceedings of the IEEE, vol. 107, no. 8, 1655-1674, 2019. DOI: https://doi.org/10.1109/JPROC.2019.2921977

Y. L. Cun, Y. Bengio, and G. Hinton. "Deep learning," Nature, vol. 521, no. 7553, pp. 436-444, 2015. DOI: https://doi.org/10.1038/nature14539

M. Horowitz, "1.1 computing's energy problem (and what we can do about it)," in IEEE international solid-state circuits conference digest of technical papers (ISSCC), 2014, pp. 10-14. DOI: https://doi.org/10.1109/ISSCC.2014.6757323

L. Lai, N. Suda, and V. Chandra. "Deep convolutional neural network inference with floating-point weights and fixed-point activations," arXiv preprint arXiv:1703.03073, 2017.

J. Wu, C. Leng, Y. Wang, Q. Hu, and J. Cheng, "Quantized convolutional neural networks for mobile devices," in IEEE conference on computer vision and pattern recognition, 2016, pp. 4820-4828. DOI: https://doi.org/10.1109/CVPR.2016.521

K. Hwang, and W. Sung. "Fixed-point feedforward deep neural network design using weights+ 1, 0, and− 1," in IEEE Workshop on Signal Processing Systems (SiPS), 2014, pp. 1-6. DOI: https://doi.org/10.1109/SiPS.2014.6986082

M. Rastegari, V. Ordonez, J. Redmon, and A. Farhadi, "Xnor-net: Imagenet classification using binary convolutional neural networks," European conference on computer vision, pp. 525-542, 2016. DOI: https://doi.org/10.1007/978-3-319-46493-0_32

M. Courbariaux, I. Hubara, D. Soudry, R. E. Yaniv, and Y. Bengio. "Binarized neural networks: Training deep neural networks with weights and activations constrained to+ 1 or-1," arXiv preprint arXiv:1602.02830, 2016.

K. A. Asha and K. D. Shinde, "Performance analysis and implementation of array multiplier using various full adder designs for DSP applications: A VLSI based approach," Intelligent Systems Technologies and Applications, pp. 731-742, 2016. DOI: https://doi.org/10.1007/978-3-319-47952-1_59

A. Nigam and R. Singh, "Comparative Analysis of 28T Full adder with 14T Full adder using 180nm," International Journal of Engineering Science Advance Research, vol. 2, no. 1, pp. 27-32, 2016.

S. Pandey, A. A. Khan, and R. Sarma. "Comparative analysis of carry select adder using 8T and 10T full adder cells," in International Conference on Communication and Signal Processing, 2014, pp. 985-989. DOI: https://doi.org/10.1109/ICCSP.2014.6949993

H. Naseri and S. Timarchi, "Low-power and fast full adder by exploring new XOR and XNOR gates," Transactions on very large scale integration (VLSI) systems, vol. 26, no. 8, pp. 1481-1493, 2018. DOI: https://doi.org/10.1109/TVLSI.2018.2820999

S. Vaidya and D. Dandekar. "Delay-power performance comparison of multipliers in VLSI circuit design," International Journal of Computer Networks & Communications (IJCNC), vol. 2, no. 4, pp. 47-56, 2010. DOI: https://doi.org/10.5121/ijcnc.2010.2405

Tải xuống

Đã Xuất bản

2025-02-28

Cách trích dẫn

[1]
Van-Khoa Pham, Lai Le, và Thanh-Kieu Tran Thi, “XNOR-Popcount, an Alternative Solution to the Accumulation Multiplication Method for Approximate Computations, to Improve Latency and Power Efficiency ”, JTE, vol 20, số p.h 01, tr 12–20, tháng 2 2025.

Số

T. 20 S. 01 (2025)

Chuyên mục

Bài báo khoa học

Categories

Giấy phép

Bản quyền (c) 2025 Tạp chí Khoa học Giáo dục Kỹ Thuật

Giấy phép Creative Commons

Tác phẩm này được cấp phép theo Giấy phép quốc tế Creative Commons Attribution-NonCommercial 4.0 .

Bản quyền thuộc về JTE.

Crossref
Scopus
Google Scholar
Europe PMC

Các bài báo được đọc nhiều nhất của cùng tác giả