Analysis of Reduced-Width Truncated Mitchell Multiplication for Inferences Using CNNs
2
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
This paper analyzes the effect of reduced output width of the truncated logarithmic multiplication and application to inferences using convolutional neural networks (CNNs). For small hardware overhead, output width is reduced in the truncated Mitchell multiplier, so that fractional bits in multiplication output are minimized in error-resilient applications. This analysis shows that when reducing output width in the truncated Mitchell multiplier, even though worst-case relative error increases, average relative error can be kept small. When adopting 8 fractional bits in multiplication output in the evaluations, there is no significant performance degradation in target CNNs compared to existing exact and original Mitchell multipliers.
Abstract
This paper analyzes the effect of reduced output width of the truncated logarithmic multiplication and application to inferences using convolutional neural networks (CNNs). For small hardware overhead, output width is reduced in the truncated Mitchell multiplier, so that fractional bits in multiplication output are minimized in error-resilient applications. This analysis shows that when reducing output width in the truncated Mitchell multiplier, even though worst-case relative error increases, average relative error can be kept small. When adopting 8 fractional bits in multiplication output in the evaluations, there is no significant performance degradation in target CNNs compared to existing exact and original Mitchell multipliers.
참고문헌 (24)
- Z. Babic, "An Iterative Mitchell’s Algorithm Based Multiplier", Proc. of ISSPIT 2008, pp.303~308 (2008)
- Z. Babi´c, "An Iterative Logarithmic Multiplier", Microprocessors and Microsystems, 35(1), pp.23~33 (2011)
- S. Hashemi, "Drum : A Dynamic Range Unbiased Multiplier for Approximate Applications", Proc. of ICCAD, pp.418~425 (2015)
- R. Zendegani, "Roba Multiplier : A Tounding-based Approximate Multiplier for High-speed yet Energy-efficient Digital Signal Processing", IEEE Trans. VLSI Systems, 25(2), pp.393~401 (2017)
- W. Liu, "Design of Approximate Radix-4Booth Multipliers for Error-tolerant Computing", IEEE Transactions on Computers, 66(8), pp.1435~1441 (2017)
- M. S. Kim, "Low-power Implementation of Mitchell’s Approximate Logarithmic Multiplication for Convolutional Neural Networks", Proc. of ASP-DAC, pp.617~622 (2018)
- I. Alouani, "A Novel Heterogeneous Approximate Multiplier for Low Power and High Performance", IEEE Embedded Systems Letters, 10(2), pp.45~48 (2018)
- J. N. Mitchell, "Computer Multiplication and Division Using Binary Logarithms", IRE Trans. Electronic Computers, 512(4), pp.512~517 (1962)
- M. S. Kim, "Efficient Mitchell’s Approximate Log Multipliers for Convolutional Neural Networks", IEEE Trans. Computers, 68(5), pp.660~675 (2018)
- S. J. Jon, "Fixed-width Multiplier for Dsp Application", Proc. of IEEE International Conference on Computer Design, pp.318~322 (2000)
- S. J. Jou, "Low-error Reduced-width Booth Multipliers for Dsp Applications", IEEE Trans. Circuits and Systems I : Fundamental Theory and Applications, 50(11), pp.1470~1474 (2003)
- K. J. Cho, "Design of Low Error Fixed-width Modified Booth Multiplier", IEEE Trans. VLSI Systems, 12(5), pp.522~531 (2004)
- S. S. Bhusare, "Fixed-width Multiplier with Simple Compensation Bias", Procedia Materials Science, 10, pp.395~402 (2015)
- K. H. Abed, "Cmos Vlsi Implementation of a Low-power Logarithmic Converter", IEEE Trans. Computers, 52(11), pp.1421~1433 (2003)
- K. Kunaraj, "Leading one Detectors and Leading one Position Detectors an Evolutionary Design Methodology", Canadian journal of electrical and computer engineering, 36(3), pp.103~110 (2013)
- S. E. Ahmed, "An Iterative Logarithmic Multiplier with Improved Precision", Proc. of IEEE Symposium Computer Arithmetic, pp.104~111 (2016)
- I. Koren, "Computer arithmetic algorithms", pp.2001, AK Peters
- Y. Jia, "Caffe : Convolutional Architecture for Fast Feature Embedding", Proc. ACM Int’l Conf. on Multimedia, pp.675~678 (2014)
- M. Lin, "Network in Network", arXiv : 1312. 4400, ~ (2013)
- Learning Multiple Layers of Features from Tiny Images, "A. Krizhevsky", Citeseer, ~ (2009)
- A. Krizhevsky, "Imagenet Classification with Deep Convolutional Neural Networks", Proc. of Advances in neural information processing systems, 1097(1105), pp.1097~1105 (2012)
- C. Szegedy, "Going Deeper with Convolutions", Proc. of the IEEE conference on computer vision and pattern recognition, pp, pp.1~9 (2015)
- K. He, "Deep Residual Learning for Image Recognition", Proc. of IEEE conference on computer vision and pattern recognition, pp.770~778 (2016)
- O. Russakovsky, "ImageNet Large Scale Visual Recognition Challenge", International Journal of Computer Vision, 115(3), pp.211~252 (2015)