Compact and Power-efficient Sobel Edge Detection with Fully Connected Cube-network-based Stochastic Computing

다국어 초록 (Multilingual Abstract)

Stochastic computing, an approximate computing method using bitstreams, has attracted attention as an alternative to deterministic computing. Stochastic computing circuits are known to perform complex calculations with high density through probability calculations. Herein, we describe the design of an accurate and compact arithmetic circuit based on stochastic computing. First, we propose a simple technique to change the output of a random number generator that is an integral part of stochastic computing for stochastic multipliers and adders. Compared with conventional designs, the results indicate that the proposed design reduces power and area and enhances the performance. This method uses a fully connected cube network and does not lose accuracy without overhead. Subsequently, when applying this design to image processing in the real world, a 63% area reduction and 95% power savings are achieved when compared to an accurate operator. Therefore, it is clear that the proposed design is optimized for energy-efficient hardware designs with high imprecision tolerance.

참고문헌 (Reference)

1 I. Sobel, "a talk at the Stanford Artificial Project" 271-272, 1968

2 I. F. Akyildiz, "Wireless mesh networks: a survey" 445-487, 2005

3 M. Yang, "Towards theoretical cost limit of stochastic number generators for stochastic computing" 154-159, 2018

4 A. Alaghi, "The promise and challenge of stochastic computing" 37 (37): 1515-1531, 2018

5 A. Alaghi, "Survey of stochastic computing" 12(2s) (12(2s)): 1-19, 2013

6 B. R. Gaines, "Stochastic computing systems" 2 : 37-172, 1969

7 A. Alaghi, "Stochastic circuits for real-time image-processing applications" 1-6, 2013

8 J. Y. F. Tong, "Reducing power by optimizing the necessary precision/range of floating-point arithmetic" 8 (8): 273-286, 2000

9 "Quartus User manual"

10 H. Joe, "Novel stochastic computing for energy-efficient image processors" 8 (8): 449-462, 2019

11 "Nangate: 45nm Open Cell Library"

12 J. Duato, "Interconnection networks" Morgan Kaufmann 2013

13 A. H. Esfahanian, "Generalized measures of fault tolerance with application to n-cube networks" 1586-1591, 1989

14 A. Alaghi, "Fast and accurate computation using stochastic circuits" 1-4, 2014

15 A. Alaghi, "Exploiting correlation in stochastic circuit design" 39-46, 2013

16 A. Sampson, "Enerj:Approximate data types for safe and general lowpower computation" 46 (46): 164-174, 2011

17 A. Alaghi, "Dimension reduction in statistical simulation of digital circuits" 1-8, 2015

18 "Design Compiler ver.M-2016.12-SP5-5"

19 "DE1-SoC User manual 1.2.2"

20 V. T. Lee, "Correlation manipulating circuits for stochastic computing" 1417-1422, 2018

21 P. Li, "Computation on Stochastic Bit Streams Digital Image Processing Case Studies" 22 (22): 449-462, 2014

22 H. Ichihara, "Compact and accurate stochastic circuits with shared random number sources" 361-366, 2014

23 A. Agrawal, "Approximate computing:Challenges and opportunities" 1-8, 2016

24 J. Han, "Approximate computing: An emerging paradigm for energyefficient design" 1-6, 2013

25 K. K. Parhi, "Analysis of stochastic logic circuits in unipolar, bipolar and hybrid formats" 1-4, 2017

26 W. Qian, "An architecture for fault-tolerant computation with stochastic logic" 65 (65): 93-105, 2011