Abstract:Multiplications have been commonly conducted in quantized CNNs, filters, and reconfigurable cores, and so on, which are widely deployed in mobile and embedded applications. Most multipliers are designed to perform multiplications with symmetric bit-widths, i.e., n - by n -bit multiplication. Such features would cause extra area overhead and performance loss when m - by n -bit multiplications ( m > n ) are deployed in the same hardware design, resulting in inefficient multiplication operations. It is highly desired and challenging to propose a reconfigurable multiplier design to accommodate operands with both symmetric and asymmetric bit-widths. In this work, we propose a reconfigurable approximate multiplier to support multiplications at various precisions, i.e., bit-widths. Unlike prior works of approximate adders assuming a uniform weight distribution with bit-wise independence, scenarios like a quantized CNN may have a centralized weight distribution and hence follow a Gaussian-like distribution with correlated adjacent bits. Thus, a new block-based approximate adder is also proposed as part of the multiplier to ensure energy-efficient operation with an awareness of the bit-wise correlation. Our experimental results show that the proposed approximate adder significantly reduces the error rate by 76% to 98% over a state-of-the-art approximate adder for Gaussian-like distribution scenarios. Evaluation results show that the proposed multiplier is 19% faster and 22% more power saving than a Xilinx multiplier IP at the same bit precision and achieves a 23.94-dB peak signal-to-noise ratio, which is comparable to the accurate one of 24.10 dB when deployed in a Gaussian filter for image processing tasks.

A Novel Multiplierless Hardware Implementation Method for Adaptive Filter Coefficients.

A Multiplier Structure Based on A Novel Real-Time Csd Recoding

A Reconfigurable Multiplier for Signed Multiplications with Asymmetric Bit-Widths.

High Quality Multiplier Design Based on Improved Circuits

Optimal Algorithm for FIR Digital Filter with Canonical Signed Digit Coefficients

A Reconfigurable Approximate Multiplier for Quantized CNN Applications.

Design of a fast multiplier used in 32Kb/s ADPCM

A hardware-effective digital decimation filter implementation for 24-bit ΔΣ ADC

An Approximating Twiddle Factor Coefficient Based Multiplier for Fixed-Point FFT

Design of low-power and high-speed multiplier

Multiplier-Free Parallel Fixed-Point Adaptive Equalizer for Real-Time Digital Coherent Communication

An Optimized Design for a Decimation Filter and Implementation for Sigma-Delta ADC

Design and Implementation of Multiplierless FIR Filter Using COOT Bird Optimization Algorithm with Different Architectures*

Low Power High Throughput Memory Less Adaptive Filter using Distributed Arithmetic

A Novel Multi-Stage Interpolation Filter Design Technique for High-Resolution ∑-△ DAC

A novel coefficient automatic calculation method for sinc filter in sigma-delta ADCs

A High-Speed, Programmable, CSD Coefficient FIR Filter

Asic Implementation of 12-Bit Radix-8 Booth Multiplier

Stochastic-Binary Hybrid Spatial Coding Multiplier for Convolutional Neural Network Accelerator

Simplified Compressor and Encoder Designs for Low-Cost Approximate Radix-4 Booth Multiplier

Efficient modular hybrid adders and Radix-4 booth multipliers for DSP applications