DEVELOPING A METHOD FOR CALCULATION AND OPTIMIZATION OF POWER CONSUMPTION OF STATIC RANDOM MEMORY DEVICES

Abstract

The relevance of minimizing the power consumption of static random access memory devices based on complementary metal-oxide-semiconductor structures is substantiated. The trend of increasing energy consumption depending on technological norms is given, from which it follows that the static power component at low technological norms grows faster and becomes commensurate with the dynamic power component. A typical block diagram of a static random access memory and its main components have been studied. An expression is given to calculate the total power consumption of static RAM. As a result of the analysis of technical literature, expressions are identified that correspond to the power consumption of the static random access memory and its main nodes, which allow estimating the distribution of power consumption between different nodes, as well as reducing power consumption at the early design stage by choosing the structural parameters of the memory device. The structure of interconnections and its delay model are considered. The corresponding expressions for calculating the resistances and capacitances of a metaloxide-semiconductor transistor and for interconnection are given. Based on the study, an optimization problem is proposed, which is essentially a non-linear programming problem. The conditions necessary for optimization are proposed, which are a system consisting of two nonlinear transcendental equations, which should be solved numerically using the MATLAB software package. To solve the problem, an approximate starting point is required, which is chosen taking into account the physical meaning of the problem. As a result of solving the problem, a local minimum point is obtained. The proposed approaches make it possible to actually solve the optimization problem.