A queueing model for evaluating the performance of parallel processing systems
Chanintorn Jittawiriyanukoon · 2006
Abstract:-This paper presents a queueing model to measure the performance of parallel processing network by introducing the 80 parallel computers for all subtasks execution. First, the case in which a task with granularity is discussed. The parallel system improves the performance by distributing and executing subtasks on dedicated 80 parallel computers. Delays in task partitioning, subtask distribution, communications and merging are taken into account for this performance evaluation. However, the prototype of this parallel environment is indeed costly. In this paper a cost-effective queueing model is proposed in order to investigate how parallel systems can function, as tasks with granularity exist. Results from cost-effective model are firstly crosschecked to results collected from 80 parallel computing units. By altering task characteristics for the mentioned parallel computing system, we can obtain the supportive results from the cost-effective queueing model. Keywords:-Parallel processing, granularity, parallelism, queueing model, performance evaluation.