Scalable and Distributed Mathematical Modeling Algorithm Design and Performance Evaluation in Heterogeneous Computing Clusters

Abstract

A growing number of scalable and distributed methods are required to effectively simulate complicated events as computing needs in the research and industrial sectors keep growing. A novel approach for developing and accessing mathematically modeled methods in heterogeneous computing clusters is proposed in this study to meet this difficulty. The suggested methodology uses DRL based Parallel Computational model for the evaluation of Heterogenous computing clusters. The algorithms makes use of parallelization methods to split up the processing burden among several nodes, supporting the variety of topologies seen in contemporary computing clusters. Through the utilization of heterogeneous hardware parts such as CPUs, GPUs, and acceleration devices, the architecture seeks to maximize speed and minimize resource usage. To evaluate the effectiveness of the proposed approach, a comprehensive performance assessment is conducted. The evaluation encompasses scalability analysis, benchmarking, and comparisons against traditional homogeneous computing setups. The research investigates the impact of algorithm design choices on the efficiency and speed achieved in diverse computing environments.