A Piecewise Linear Mitchell Algorithm-Based Approximate Multiplier

In the field of integrated circuits, computational cost has always been a crucial design metric. recent years, with continuous development in computing, requirements for computation have growing rapidly. Reducing and improving efficiency become key issues field. There are many error-tolerant applications multimedia where approximate computing techniques can be applied to improve reduce costs at acceptable errors. This paper proposed piecewise linear Mitchell algorithm based on logarithmic approximation multiplication algorithm. Additionally, Pwl-Mit multiplier is designed according improved combined data truncation technique. The better statistical performance compared state-of-the-art multipliers. simulated synthesized TSMC 65 nm process, its reliability verified using discrete cosine transform (DCT) transform.