Towards an adaptive OS noise mitigation technique for microbenchmarking on mobile platforms

This study investigates levels of Operating System (OS) noise on Apple iPad mobile devices. OS noise causes variations in application performance that interfere with microbenchmark results. OS noise manifests in collected data through extreme outliers and variations in skewness. Using our collected data, we develop an iterative, semi-automated outlier removal process for OS noise profiles. The profiles generated by outlier removal represent the first step toward an adaptive noise mitigation technique for use in microbenchmarking for mobile platforms