Approximate computing is a promising technique to elevate the performance of digital circuits at the cost of reduced accuracy in numerous error-resilient applications. Multipliers play a key role in many of these applications. In this brief, we propose a truncation based Booth multiplier with a compensation circuit generated by selective modifications in k-map to circumvent the carry appearing from the truncated part. By judicious mapping, hardware pruning and output error reduction is achieved simultaneously. In the quest of power and accuracy trade-off, Truncated and Approximate Carry based Booth Multipliers (TACBM) are proposed with a range of designs based on truncation factor w . When compared with the state-of-the-art multipliers, TACBM outperforms in terms of accuracy and Area-Power savings. TACBM( w=10 ) provides with 0.02% MRED and 23% reduction in Area-Power product compared to exact Booth multiplier. The multipliers are evaluated using image blending and Multilayer perceptron (MLP) neural network and a high value of accuracy (95.63%) for MLP is achieved.
Software Implementation:
Modelsim
Xilinx
” Thanks for Visit this project Pages – Register This Project and Buy soon with Novelty “
Area and Power Efficient Truncated Booth Multipliers Using Approximate Carry-Based Error Compensation
