Simulation modelling of AI clusters

Abstract

This project leverages CQSim, a cloud queue-based scheduling simulator, to implement and evaluate scheduling algorithms tailored for heterogeneous compute resources. The project integrates and modifies scheduling policies to analyze their impact on resource utilization, job completion time, and fairness. The findings highlight the importance of workload-aware scheduling in heterogeneous clusters, demonstrating how workload-aware strategies enhance system efficiency. This study provides insights into optimizing scheduling policies for next-generation data centers managing AI and high-performance computing workloads.