Dynamic accuracy adjustement for fixed width dividers

Divisions are needed in several signal processing applications and they are widely destined to hardware implementations. However, they often constitute the performance bottleneck because of the poor performances of the related large hardware dividers. We present a generic VLSI design to respond to some applications need of reducing dividers input width. Unlike any other published method, in our solution, the word's width is dynamically adjusted while indicating which level of accuracy that is being processed. FPGA implementation results are provided to show our design's low area and timing consumption to recover information that is always lost when usual Least Significant Bit (LSB) reduction is applied.

[1]  S. Tagzout,et al.  Compact parallel multipliers using the sign-generate method in FPGA , 1997, 1997 21st International Conference on Microelectronics. Proceedings.

[2]  Javier Valls-Coquillat,et al.  FPGA-implementation of atan(Y/X) based on logarithmic transformation and LUT-based techniques , 2010, J. Syst. Archit..

[3]  Adel Belouchrani,et al.  Arctangent Architecture for High Speed and High Precision Data , 2011, J. Circuits Syst. Comput..

[4]  Randy Yates,et al.  Fixed-Point Arithmetic: An Introduction , 2013 .

[5]  Dong-U Lee,et al.  A Bit-Width Optimization Methodology for Polynomial-Based Function Evaluation , 2007, IEEE Transactions on Computers.

[6]  Yajun Ha,et al.  An automated, efficient and static bit-width optimization methodology towards maximum bit-width-to-error tradeoff with affine arithmetic model , 2006, Asia and South Pacific Conference on Design Automation, 2006..

[7]  Arnaud Tisserand,et al.  Multipartite table methods , 2005, IEEE Transactions on Computers.

[8]  Wayne Luk,et al.  Optimizing hardware function evaluation , 2005, IEEE Transactions on Computers.