Brawny vs. Wimpy: Evaluation and Analysis of Modern Workloads on Heterogeneous Processors

Heterogeneous processors, consisting of a mix of high-performance 'brawny' processors and low-power 'wimpy' processors, have been proposed to achieve higher energy-efficiency by making it possible for different applications within a diverse mix of workloads to be run on the most appropriate cores. This paper performs a comparative analysis of such heterogeneous multicores on the performance and energy efficiency of mobile devices and server systems. Using several real-world workloads from both mobile and enterprise domain, experimental evaluations are carried out on a unique experimental test bed comprised of real heterogeneous CPUs that differ in both their core architecture and cache sizes, comparing the performance and efficiency for these applications. Results show that heterogeneous processors can improve the performance and energy-efficiency for these diverse applications. They also highlight the challenges associated with user-perceived performance on mobile devices and power overhead of shared system components on the energy-efficiency of heterogeneous cores and thus call for novel energy-aware resource management approaches. The experimental results presented in this paper provide platform and system software designers a perspective on the trade-offs involved with these architectures and thus make optimal design choices.