Modern data centers become increasingly power limited, due to continuously increasing demand on cloud resources. As achieving low power and high performance are competing objectives, the efficient management of power/performance tradeoffs is one of the most challenging tasks in modern computing systems. In order to address the increasing management complexity, various types of adaptive techniques, such as system-level DVFS or application-level modifications have been proposed previously. In this work, we propose and implement an adaptive framework that jointly utilizes system and application-level adaptation to improve efficiency of multi-core servers. Our results show that joint- management reduces the power consumption up to 27% and improves the performance up to 2.7x under power and performance constraints.