Enhanced bit-width optimization for linear circuits with feedbacks

Bit widths to accuracy trade offs require complex method development and testing. This paper presents a new process of incremental optimizations to get to reduced bit-widths compared to recently published results. Range and Precision improvements reach all the data path interconnections. More, a precision refinement is proposed even for input data. Detailed procedures and results of hardware implementations are provided to show the achieved improvements compared to existing optimization approaches. For illustration and to compare our work to existing results, different configurations of an IIR filter is used for tests and implementations. Depending on those different configurations, the proposed optimization techniques present an average of area reductions ranging from 17% to 30%.

[1]  O. Sarbishei,et al.  Analysis of range and precision for fixed-point linear arithmetic circuits with feedbacks , 2010, 2010 IEEE International High Level Design Validation and Test Workshop (HLDVT).

[2]  Leland B. Jackson,et al.  On the interaction of roundoff noise and dynamic range in digital filters , 1970, Bell Syst. Tech. J..

[3]  Katarzyna Radecka,et al.  Using arithmetic transform for verification of datapath circuits via error modeling , 2000, Proceedings 18th IEEE VLSI Test Symposium.

[4]  Katarzyna Radecka,et al.  An efficient hybrid engine to perform range analysis and allocate integer bit-widths for arithmetic circuits , 2011, 16th Asia and South Pacific Design Automation Conference (ASP-DAC 2011).

[5]  Robert W. Brodersen,et al.  Automated fixed-point data-type optimization tool for signal processing and communication systems , 2004, Proceedings. 41st Design Automation Conference, 2004..

[6]  Joan Carletta,et al.  Determining appropriate precisions for signals in fixed-point IIR filters , 2003, Proceedings 2003. Design Automation Conference (IEEE Cat. No.03CH37451).

[7]  Romuald Rocher,et al.  Analytical Fixed-Point Accuracy Evaluation in Linear Time-Invariant Systems , 2008, IEEE Transactions on Circuits and Systems I: Regular Papers.

[8]  Wayne Luk,et al.  Accuracy-Guaranteed Bit-Width Optimization , 2006, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[9]  Wayne Luk,et al.  MiniBit: bit-width optimization via affine arithmetic , 2005, Proceedings. 42nd Design Automation Conference, 2005..

[10]  Katarzyna Radecka,et al.  Arithmetic Transforms of Imprecise Datapaths by Taylor Series Conversion , 2006, 2006 13th IEEE International Conference on Electronics, Circuits and Systems.

[11]  A. Oppenheim,et al.  Effects of finite register length in digital filtering and the fast Fourier transform , 1972 .

[12]  Nicola Nicolici,et al.  Bit-Width Allocation for Hardware Accelerators for Scientific Computing Using SAT-Modulo Theory , 2010, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[13]  Wayne Luk,et al.  Automatic Accuracy-Guaranteed Bit-Width Optimization for Fixed and Floating-Point Systems , 2007, 2007 International Conference on Field Programmable Logic and Applications.

[14]  J. M. Pierre Langlois,et al.  Enhanced Precision Analysis for Accuracy-Aware Bit-Width Optimization Using Affine Arithmetic , 2013, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[15]  Gabriel Caffarena,et al.  SQNR Estimation of Fixed-Point DSP Algorithms , 2010, EURASIP J. Adv. Signal Process..

[16]  O. Sarbishei,et al.  Analytical Optimization of Bit-Widths in Fixed-Point LTI Systems , 2012, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[17]  Wonyong Sung,et al.  Combined word-length optimization and high-level synthesis ofdigital signal processing systems , 2001, IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..