作者: Ravi Iyer , Li Zhao , Fei Guo , Ramesh Illikkal , Srihari Makineni
关键词:
摘要: As we enter the era of CMP platforms with multiple threads/cores on die, diversity simultaneous workloads running them is expected to increase. The rapid deployment virtualization as a means consolidate single platform prime example this trend. In such scenarios, quality service (QoS) that each individual workload gets from can widely vary depending behavior simultaneously workloads. While number cores assigned be controlled, there no hardware or software support in today's control allocation resources cache space and memory bandwidth paper, propose QoS-enabled architecture for addresses problem. enables more (i.e. space) bandwidth) high priority applications based guidance operating environment. also allows dynamic resource reassignment during run-time further optimize performance application minimal degradation low priority. To achieve these goals, will describe hardware/software required well environment (O/S virtual machine monitor). Our evaluation framework consists detailed simulation models version Linux. Based experiments, show effectiveness summarize key findings/trade-offs.