In an interactive operating system, the quest for ensuring timely response to user input has been a challenge. Differentiating interactive tasks from batch tasks were the first segregation to ensure lower latency in responding to users' inputs. The introduction of different task priority levels further enhances this. As hardware performance improves, users expect immediate response to any interaction and hence operating system developers have been turning to the scheduler algorithm to ensure good interactivity. This paper presents a study on interactivity performance for three different scheduling algorithms that are currently available as part of different Linux distributions. The performances were measured using the Interbench tool where the interactivity measurements were then used to support the scheduling latencies that were attributed to the scheduling algorithm. In addition, this paper also produced new findings on the schedulers' performances in the presence of different load priority levels.
展开▼
