Authors: Glen Gilchrist
Comments: 10 Pages.
Obtaining additional computational speed from a central processing unit by means of driving a computer system at clock frequencies higher than the default settings has long been used as a method to inexpensively “boost” the performance of a computer. With the emergence of so called smart-phones and the openness of the Android operating system, such tweaks have recently been applied to mobile handsets. This paper investigates the performance gains to an off the shelf handset with the custom Skatie Rom (C3C0, 2012) via means of a statistically designed experiment. A Taguchi Orthogonal Array was used to investigate 5 factors, each at 3 levels on the performance as measured with Aurora Softworks Quadrant application. Unsurprisingly the core CPU speed had the largest effect on overall performance, but we demonstrate that CPU Governor (lagfree) and the I/O Scheduler (noop) were also significant at p=0.000, whilst the size of the SD Card Cache is significant to p=0.065.
In the present study, we propose estimators based on geometric and harmonic mean
for estimating population mean using information on two auxiliary attributes in simple
random sampling. We have shown that, when we have multi-auxiliary attributes, estimators
based on geometric mean and harmonic mean are less biased than Olkin (1958), Naik and
Gupta (1996) and Singh (1967) type- estimator under certain conditions. However, the MSE
of Olkin( 1958) estimator and geometric and harmonic estimators are same up to the first
order of approximation.