Exploring Practical Benefits of Asymmetric Multicore Processors

Asymmetric multicore processors (AMP) are built of cores that expose the same ISA but differ in performance, complexity, and power consumption. A typical AMP might consist of a plenty of slow, small and simple cores and a handful of fast, large and complex cores. AMPs have been proposed as a more energy efficient alternative to symmetric multicore processors. They are particularly interesting in their potential to mitigate Amdahl’s law for parallel program with sequential phases. While a parallel phase of the code runs on plentiful slow cores enjoying low energy per instruction, the sequential phase can run on the fast core, enjoying high single-thread performance of that core. As a result, performance per unit of energy is maximized. In this paper we evaluate the effects of accelerating sequential phases of parallel applications on an AMP. Using a synthetic workload generator and an efficient asymmetry-aware user-level scheduler, we explore how the workload’s properties determine the speedup that the workload will experience on an AMP system. Such an evaluation has been performed before only analytically; experimental studies have been limited to a small number of workloads. Our study is the first to experimentally explore benefits on AMP systems for a wide range of workloads.