A High-Efficiency General Nonlinear Function Generator for Stochastic Computation
Xincheng Feng, Ke Ping Hu, Kaining Han · 2021
Nonlinear function calculation is widely used in numerous science and technology fields. Stochastic calculation has great advantages in reducing computational complexity and improving hardware efficiency. However, stochastic calculation has the bottleneck of slow convergence speed, low calculation accuracy and poor generalization. In this paper, we propose a multiple driving and multiple dimension finite state machine (MM-FSM) to implement major single variable nonlinear functions used in information and signal processing areas on with low complexity, low latency, and considerable generalization. We will provide the corresponding synthesis method of the activation parameters and conditional parameters of MM-FSM. Compared with the traditional stochastic scheme and Coordinate Rotation Digital Computer (CORDIC) algorithm, simulation results show that the proposed MM-FSM nonlinear function generator has significantly lower complexity while guaranteeing the calculation accuracy.