A number of techniques have been proposed to provide runtime performance guarantees while minimizing power consumption. However, existing approaches typically operate on a fixed set of components (or actuators) specified at design time. This limitation requires redesign and reimplementation if new components become available.

In this paper, we introduce PTRADE, a novel performance management framework that is general with respect to the components it manages. PTRADE can be deployed on new systems with different components without requiring redesign or reimplementation. The framework’s flexibility is demonstrated through its management of performance goals across a variety of benchmarks on two different Linux/x86 systems and a simulated 128-core system, each featuring distinct components that influence power and performance tradeoffs.

Our experimental results indicate that PTRADE achieves its generality while meeting performance goals with low error and achieving near-optimal power consumption. This work highlights PTRADE as a versatile and efficient solution for adaptive performance management in diverse computing environments.