Key points are not available for this paper at this time.
Composite Cores Architecture (CCA), a class of dynamic heterogeneous architectures, enables the system to construct the right core at run-time for each application by composing cores together to build larger core or decomposing a large core into multiple smaller cores. While this architecture provides more flexibility for the running application to find the best run-time settings to maximize energy-efficiency, due to interdependence of various tuning parameters such as the type of the core, run-time voltage and frequency and the number of threads, it makes it more challenging for scheduling. Past research mainly addressed the scheduling problem in composite cores architecture by looking at one or two of these tuning parameters. However, as we will show in this paper, it is important to concurrently optimize and fine-tune these parameters to harness the power of heterogeneity in this emerging class of architecture. In addition, most previous work on CCA mainly studied traditional single threaded CPU applications. In this work, we investigate the scheduling challenges for multithreaded applications in CCA. First, through methodical investigation of power and performance results, we characterize various multithreaded applications on a CCA which can be composed into few big or many little cores and demonstrate how the interplay among various application, system, and architecture level parameters affect the performance and energy-efficiency. Furthermore, based on characterization results, a highly accurate regression-based model for energy-efficiency prediction is developed to guide the scheduling decision. Using the predictive model, we developed a scheduling scheme for effective mapping of multithreaded applications onto CCA. The results show that the proposed scheduling scheme on average achieves close to 94% efficiency as compared to the Oracle scheduling.
Sayadi et al. (Sun,) studied this question.