USING DLP TO OPTIMIZE COMPUTER MODELS IN SCADA SYSTEMS

This paper presents a study on the efficiency of applying data-level parallelism using SIMD technology to optimize computer models within SCADA systems. The focus is on implementing computational algorithms based on explicit and implicit finite difference schemes, which are widely used in industrial process modeling. Experimental research demonstrates improved computational performance achieved through the use of modern SIMD instructions, primarily AVX-256, enabling real-time data processing for complex simulation models in industrial control systems. It was found that explicit finite difference schemes exhibit better suitability for vectorization compared to implicit schemes implemented using the Thomas algorithm. The practical significance of the results lies in enhancing the efficiency of SCADA systems through the optimal utilization of modern CPU capabilities, offering tangible benefits for industrial enterprises adopting this technology.