ABSTRACT
Approximate computing has been recognized as an effective low power technique for applications with intrinsic error tolerance, such as image processing and machine learning. Existing efforts on this front are mostly focused on approximate circuit design, approximate logic synthesis or processor architecture approximation techniques. This work aims at how to make good use of approximate circuits at system and block level. In particular, approximation aware scheduling, functional unit allocation and binding algorithms are developed for data intensive applications. Simple yet credible error models, which are essential for precision control in the optimizations, are investigated. The algorithms are further extended to include bitwidth optimization in fixed point computations. Experimental results, including those from Verilog simulations, indicate that the proposed techniques facilitate desired energy savings under latency and accuracy constraints.
- V. Gupta, D. Mohapatra, S. P. Park, A. Raghunathan, and K. Roy. IMPACT: imprecise adders for low-power approximate computing. ISLPED, pages 409--414, 2011. Google ScholarDigital Library
- J. Miao, K. He, A. Gerstlauer, and M. Orshansky. Modeling and synthesis of quality-energy optimal approximate adders. ICCAD, pages 728--735, 2012. Google ScholarDigital Library
- J. Miao He, A. Gerstlauer, and M. Orshansky. Approximate logic synthesis under general error magnitude and frequency constraints. ICCAD, pages 779--786, 2013. Google ScholarDigital Library
- C. Liu, J. Han, and F. Lombardi. A low-power, high-performance approximate multiplier with configurable partial error recovery. DATE, 2014. Google ScholarDigital Library
- K. Nepal, Y. Li, R. I. Bahar, and S. Reda. ABACUS: a technique for automated behavioral synthesis of approximate computing. DATE, 2014. Google ScholarDigital Library
- S. Venkataramani, A. Ranjan, K. Roy, and A. Raghunathan. AxNN: energy-efficient neuromorphic systems using approximate computing. ISLPED, pages 27--32, 2014. Google ScholarDigital Library
- R. Venkatesan, A. Agarwal, K. Roy, and A. Raghunathan. MACACO: modeling and analysis of circuits for approximate computing. ICCAD, pages 667--673, 2011. Google ScholarDigital Library
- J. Huang, J. Lach, and G. Robins. Analytic error modeling for imprecise arithmetic circuits. SELSE, 2011.Google Scholar
- J. Liang, J. Han, and F. Lombardi. New metrics for the reliability of approximate and probabilistic adders. IEEE Trans. on Computers, 62(9):1760--1771, June 2012. Google ScholarDigital Library
- W.-T. J. Chan, A. B. Kahng, S. Kang, R. Kumar, and J. Sartori. Statistical analysis and modeling for error composition in approximate computation circuits. ICCD, pages 47--53, 2013.Google ScholarCross Ref
- D.-U. Lee, A. A. Gaffar, R. C. C. Cheung, O. Mencer, W. Luk, and G. A. Constantinides. Accuracy-guaranteed bit-width optimization. IEEE TCAD, 25(10):1990--2000, October 2006. Google ScholarDigital Library
- S. Lee and A. Gerstlauer. Fine graind word length optimization for dynamic precision scaling in DSP systems. VLSI-SoC, pages 266--271, 2013.Google Scholar
- K.-I. Kum and W. Sung. Combined word-length optimization and high-level synthesis of digital signal processing systems. IEEE TCAD, 20(8):921--930, August 2001. Google ScholarDigital Library
- J. Cong, Y. Fan, G. Han, Y. Lin, J. Xu, Z. Zhang, and X. Cheng. Bitwidth-aware scheduling and binding in high-level synthesis. ASPDAC, pages 856--861, 2005. Google ScholarDigital Library
- C. Lee, M. Potkonjak, and W. H. Mangione-Smith. MediaBench: a tool for evaluating and synthesizing multimedia and communications systems. MICRO, pages 330--335, 1997. Google ScholarDigital Library
- IBM. CPLEX Optimizer. http://www-01.ibm.com/software/commerce/optimization/cplex-optimizer/, 2014.Google Scholar
Index Terms
- Joint precision optimization and high level synthesis for approximate computing
Recommendations
Layout-driven RTL binding techniques for high-level synthesis
ISSS '96: Proceedings of the 9th international symposium on System synthesisThe importance of effective and efficient accounting of layout effects is well-established in high-level synthesis (HLS), since it allows more realistic exploration of the design space and the generation of solutions with predictable metrics. This ...
A high-level register optimization technique for minimizing leakage and dynamic power
GLSVLSI '07: Proceedings of the 17th ACM Great Lakes symposium on VLSIA large fraction of the total power dissipated in a digital circuit is consumed by the clocked elements in the data-path. Hence, savings in power usage of these components can be directly reflected in a circuit's overall power consumption. Reducing ...
High-level synthesis for easy testability
EDTC '95: Proceedings of the 1995 European conference on Design and TestThis paper presents an attempt towards design quality improvement by incorporation of testability features during datapath high-level synthesis. This method is based on the use of hardware sharing possibilities to improve the testability of the circuit ...
Comments