No abstract

Abstract-Runtime power management using dynamic voltage and frequency scaling (DVFS) has been extensively studied for video processing applications. But there is little work on game power management although gaming applications are now widely run on battery-operated portable devices like mobile phones. Taking a cue from video power management, where PID controllers have been successfully used, they were recently applied to game workload prediction and DVFS. However, the use of hand-tuned PID controller gains on relatively short game plays left open questions on the robustness of the controller and the sensitivity of prediction quality on the choice of the gain values. In this paper we try to systematically answer these questions. We first show that from the space of PID controller gain values, only a small subset leads to good game quality and power savings. Further, the choice of this set highly depends on the scene and the game application. For most gain values the controller becomes unstable, which can lead to large oscillations in the processor's frequency setting and thereby poor results. We then study a number of time series models, such as a Least Mean Squares (LMS) Linear Predictor and its generalizations in the form of Autoregressive Moving Average (ARMA) models. These models learn most of the relevant model parameters iteratively as the game progresses, thereby dramatically reducing the complexity of manual parameter estimation. This makes them deployable in real setups, where all game plays and even game applications are not a priori known. We have evaluated each of these models (PID, LMS and ARMA) for a variety of games -ranging from Quake II to more recent closed-source games such as Crysis, Need for Speed -Shift and World in Conflict -with very encouraging results. To the best of our knowledge, this is the first work that systematically explores (a) the feasibility of manually tuning PID controller parameters for power management, (b) time series models for workload prediction for gaming applications, and (c) power management for closed-source games.

While dynamic voltage and frequency scaling (DVFS) based power management has been widely studied for video processing, there is very little work on game power management. Recent work on proportional-integral-derivative (PID) controllers for predicting game workload used hand tuned PID controller gains on relatively short game plays. This left open questions on the robustness of the PID controller and how sensitive the prediction quality is on the choice of the gain values, especially for long game plays involving different scenarios and scene changes. In this paper we propose a Least Mean Squares (LMS) Linear Predictor, which is a regression model commonly used for system parameter identification. Our results show that game workload variation can be estimated using a linear-in-parameters (LIP) model. This observation dramatically reduces the complexity of parameter estimation as the LMS Linear Predictor learns the relevant parameters of the model iteratively as the game progresses. The only parameter to be tuned by the system designer is the learning rate, which is relatively straightforward. Our experimental resnlts nsing the LMS Linear Predictor show comparable power savings and game quality with those obtained from a highly-tuned PID controller. I. INTRODU CTIONGraphics-intensive game applications gained significant popularity in recent years. Although most of them are available on high-end desktops, the advent of these appli cations on battery-powered mobile devices (e.g., laptops, PDAs, cell phones and portable game consoles) is steadily increasing. This recent development is resulting in a con stantly widening gap between the demand for computational resources on portable devices and the corresponding energy resources available through batteries [2]. In this context, power management techniques play a significant role in reducing this gap by increasing the energy efficiency of these devices. Most of these devices are equipped with dynamic voltage/frequency-scalable processors in which the power dissipation per clock cycle is directly proportional to its frequency and the square of the supply voltage. There fore, one can reduce energy consumption through dynamic voltage/frequency scaling techniques, where the processor's clock frequency is dynamically adjusted in response to a varying workload.Over the last few years, DVFS based power manage ment schemes have been widely explored and successfully applied to audio [3], digital signal processing, and video frame decoding/encoding applications [1], [5], [12], [17]. However, the development of these schemes in the domain 978-1-4244-8935-0/10/$26.00 ©201 0 IEEE of interactive game applications is still in its infancy. This is mainly due to the fact that game applications are highly interactive in nature, where the content is dynamically generated, making it impossible to buffer frames, as it is done in the case of audio or video processing applications.In this context, recent work [ 10] has shown that game frames exhibit sufficient workload variability, making g...

Abstract-Power consumption and battery life are important design concerns for mobile platforms. On these devices games can be considered as one of the most demanding applications in terms of computational cost and consumed energy. In this demo we showcase Android-based power management for games. We reduce the power consumption of games by scaling the processor's voltage and frequency. Towards this, the game's future workload has to be predicted. To accurately predict the workload, previous work heavily instrumented the game's source code itself. The source code is typically not available for up-todate Android games. The work presented in this paper does not require any modification of the game's source code and therefore can as well be applied to closed source games. Towards this, we utilize the game's communication interfaces with the operating system to accurately predict a game's workload. The approach presented in the following has been implemented and tested on the PandaBoard ES [12] and Galaxy Nexus mobile phone with a number of popular closed-source games. Measurements show significant power savings while the gaming experience is maintained.

No abstract