On Algorithms of Task Placement in Computing Clusters Taking into Account Task and System Topologies

This paper considers various task placement algorithms for parallel computing tasks on clusters that take into account or disregard the task and cluster topology. Clusters with torus and tree topologies are simulated using the SimGrid framework. Experiments conducted on these models evaluate the effectiveness of the algorithms. The results indicate that using task placement algorithms considering both the system and task topology significantly accelerate task execution.