Optimizing Windows Vista Platforms for Energy Efficiency

Optimizing Windows Vista Platforms for Energy Efficiency - 7

Optimizing Windows Vista Platforms for Energy Efficiency

May 1, 2007

Abstract

This paper enables system manufacturers and IT professionals to configure Windows Vista™ platforms for maximum energy efficiency. The techniques in this paper may be used to help extend mobile PC battery life and reduce energy operation expenses for desktop and server PCs. This paper explains how to evaluate system energy efficiency and demonstrates example power policy settings to favor power savings or performance. Best practices for energy-efficient platform design are also covered.

This information applies for the Windows Vista operating system.

The current version of this paper is maintained on the Web at:
http://www.microsoft.com/whdc/system/pnppwr/powermgmt/optimize_Power.mspx

References and resources discussed here are listed at the end of this paper.

Contents

Introduction 3

Preliminary Reading 3

Evaluating Platform Energy Efficiency 3

Evaluation Scenarios and Workloads 4

Measuring Average System and Device Power Consumption 4

Measuring Utility Power 5

Measuring Mobile PC Battery Life 5

Indirect Indicators of System Power Consumption 6

Configuring Windows Vista Power Policy for Maximum Energy Efficiency 7

Display Idle Time-Out 7

Sleep Idle Time-Out 8

Hard Disk Drive Idle Time-Out 8

Display Brightness 9

Processor Power Management 9

802.11 Wireless Power Save Mode 10

PCI Express Active-State Power Management 11

USB Selective Suspend 11

Windows Search and Indexing 12

Idle Detection Threshold 12

SATA AHCI Link Power Management 13

Power Policy Changes for Sleep Energy Efficiency 14

Wake on Timer 14

USB Host Controller Device Power State 15

Platform Design Guidelines for Energy Efficiency 16

Evaluate Preinstallation Software for Energy Efficiency 16

Efficient Use of USB Devices and Selective Suspend 16

Conclusion 17

Call to Action 17

References 18

Additional Resources 18

Disclaimer

This is a preliminary document and may be changed substantially prior to final commercial release of the software described herein.

The information contained in this document represents the current view of Microsoft Corporation on the issues discussed as of the date of publication. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information presented after the date of publication.

This White Paper is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS DOCUMENT.

Complying with all applicable copyright laws is the responsibility of the user. Without limiting the rights under copyright, no part of this document may be reproduced, stored in or introduced into a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), or for any purpose, without the express written permission of Microsoft Corporation.

Microsoft may have patents, patent applications, trademarks, copyrights, or other intellectual property rights covering subject matter in this document. Except as expressly provided in any written license agreement from Microsoft, the furnishing of this document does not give you any license to these patents, trademarks, copyrights, or other intellectual property.

Unless otherwise noted, the example companies, organizations, products, domain names, e-mail addresses, logos, people, places and events depicted herein are fictitious, and no association with any real company, organization, product, domain name, email address, logo, person, place or event is intended or should be inferred.

© 2007 Microsoft Corporation. All rights reserved.

Microsoft, Windows, and Windows Vista are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries.

The names of actual companies and products mentioned herein may be the trademarks of their respective owners.

Introduction

Energy efficiency is currently one of the most important topics in the computing industry. Processor and chipset manufacturers are promoting their hardware in terms of performance per watt consumed. Governmental agencies are increasing their scrutiny on PC energy consumption and encouraging consumers to purchase energy-efficient hardware and software.

Energy efficiency impacts all Microsoft® Windows® platforms from ultra-mobile systems to many-processor servers. Energy efficiency helps extend battery life in mobile PCs and helps reduce energy expenses for desktop PCs and server systems.

Windows Vista™ helps enable energy efficiency on PC platforms by providing power policy configuration options. System manufacturers and IT professionals may use power policy configuration options to tune Windows Vista platforms for power savings or performance. Simple changes to power policy configuration can help extend mobile PC battery life or enable extra system processor and disk performance.

Windows Vista also provides group policy support that allows IT professionals to enforce power policy settings across an enterprise. This functionality can be used to enable monitor and computer power management settings to help reduce PC and server energy expenses.

This paper demonstrates the methods to configure Windows Vista platforms for maximum energy efficiency. System manufacturers will learn how to evaluate platform energy and then customize Windows Vista power policy for their hardware platform. IT professionals will learn detailed information about many Windows Vista power policy options that can be used to help reduce energy expenses in the enterprise. Finally, several best practices for hardware platform design are discussed.

Preliminary Reading

Readers of this paper are encouraged to first review “Power Policy Configuration and Deployment for Windows Vista,” which details the Windows Vista power policy configuration architecture. The paper also gives a brief overview of group policy and the PowerCfg command-line utility that can be used to change detailed system power policies.

Evaluating Platform Energy Efficiency

To properly optimize Windows Vista platforms for energy efficiency, it is important to have a benchmark and baseline with which to compare platform adjustments and improvements. The methods to evaluate energy efficiency vary by platform and audience. For example, mobile PC system manufacturers typically evaluate battery life by using a battery run-down test or an industry-standard benchmark. IT professionals may measure desktop and server power consumption by using a utility power meter or server power distribution hardware.

Evaluation Scenarios and Workloads

Energy efficiency requires evaluation of power consumption over a specific workload or scenario. The workload varies by platform and task, and may be one of:

·  Idle
The system is running without applications executing and is only processing Windows background tasks. The monitor/display device may or may not be enabled, depending on the desired evaluation scenario. Energy efficiency during the idle scenario is critical because a desktop or mobile PC is most often idle when it is powered on.

·  DVD playback
A typical consumer scenario—a DVD title or chapter is played repetitively to evaluate the power consumption of DVD playback on a given system.

·  Productivity workload
A productivity workload evaluates energy efficiency by simulating a user interacting with the system through knowledge-worker scenarios, such as use of Microsoft Office applications. Productivity workloads typically also have intermittent periods of idle and task switching to simulate real user behavior.

Energy-efficiency evaluation workloads differ from other PC benchmarking scenarios in that they are typically not designed to exercise maximum performance from the system.

System manufacturers that include value-added software to their Windows preinstallation images should evaluate energy efficiency with all value-added software installed and enabled. This ensures that the evaluation closely matches the end-user experience with the system.

Measuring Average System and Device Power Consumption

A common method to evaluate system energy efficiency is to measure average system or device power consumption over a given workload. This form of energy efficiency measurement is typically used with instrumented hardware and therefore most appropriate for system manufacturers and processor or chipset vendors.

Instrumented hardware is designed to allow power consumption measurement of one or more system hardware components while the system is in use. Instrumented systems require special measurement hardware and software components to average power consumption over the workload.

The average system or device power consumption may be compared before and after software or hardware changes. The differences in average power indicate a positive or negative change in energy efficiency over the same workload.

Average system power consumption may also be measured on mobile PCs by measuring power at the system DC input. Most mobile PCs have a separate AC-to-DC power converter that connects between the mobile PC chassis and utility power. Measuring power consumption at the DC input is preferred over the AC/utility power connection because power conversion losses are not measured.

Measuring Utility Power

The average power consumption of a desktop or server system may be measured where it connects to utility power. This energy-evaluation technique is appropriate for IT professionals attempting to evaluate the energy efficiency of a desktop or server system and the impact of monitor or computer power management settings.

This energy-evaluation technique is appropriate for systems that cannot be instrumented or systems that are already installed and in use.

Many devices are available to measure the utility power consumed by a desktop or server system and its displays. Typically, these devices connect between the system and their utility power connection. In rack-mount server scenarios, the power distribution or uninterruptible power supply (UPS) equipment may already have power consumption measurement capabilities. Utility power measurement devices can often store and average power consumption over several hours, days, or weeks.

Measuring utility power is not as accurate as evaluating energy efficiency with an instrumented system and therefore is not appropriate for measuring the impact of subtle driver or software changes.

Measuring Mobile PC Battery Life

The most common energy-efficiency evaluation is determining mobile PC battery life.

System manufacturers are encouraged to evaluate the battery life of their Windows Vista mobile PCs. The suggested process for this evaluation is:

1. Determine the evaluation scenario: Idle, DVD Playback, Productivity or Knowledge-worker scenario.

2. Install Windows Vista and minimum driver components.

Install Windows Vista on the system and all device drivers. Ensure that no uninstalled device drivers or other errors are present in Device Manager. Do not install value-added software such as additional communications or anti-virus utilities.

3. Evaluate battery life with a run-down test.

Fully charge the system battery and allow the system to execute the scenario (repetitively if necessary) until the battery has fully drained. Disable the system Sleep and Hibernate idle timers to ensure that the battery drains fully to empty. Track the amount of time that the system was able to execute the workload before powering off due to the empty battery. It is critical to repeat this several times and average the workload task duration to account for any variance in battery drain times.

4. Repeat the run-down test with value-added applications and services installed.

To effectively gauge the impact of preinstallation applications on battery life, it is recommended to repeat the run-down test with value-added applications installed and enabled. The run-down test should be repeated multiple times and the battery life duration averaged to account for any slight variance in battery drain times between runs.

Several methods may be employed to determine the battery run-down duration time. One method is to periodically write a timestamp to a file on the disk. However, this method prevents aggressive disk spin-down policies from working correctly. Another available option is to monitor the system under test externally via Ethernet link status or display status to determine when the battery run-down test has ended.

Battery devices are impacted by many variables that affect run-down test durations including ambient temperature, time to full charge, and time between runs. System manufacturers performing battery life run-down tests are encouraged to average the run-down durations over three or more runs for a given workload.

Indirect Indicators of System Power Consumption

Overall system power consumption and relative energy efficiency may be estimated indirectly by using several other performance and power-saving indicators. For example, mobile PCs do not realize their maximum battery life time if processor power management features are not working correctly.

System manufacturers are encouraged to validate correct operation of the following platform power management features to help ensure maximum energy efficiency:

·  Processor performance states
If the platform processor is capable of dynamic frequency and voltage scaling, ensure that this capability is exposed and properly working within Windows Vista. Use the PwrTest utility—available in the Windows Driver Kit (WDK)—or Performance Monitor to monitor the use of processor performance states.

·  Processor idle states
If the platform supports processor idle states, validate effective operation within Windows Vista. A common problem with efficient operation of processor idle states is USB devices that do not support USB selective suspend.

·  System timer tick interval
Ensure that the platform timer tick interval is set to the largest value possible (typically, 15.6 ms). A smaller platform timer tick interval prevents effective use of processor idle states, increasing overall system power consumption. The Windows Event Viewer can be used to view system timer tick change events in the Microsoft-Windows-Kernel-Power\Diagnostic log.

Configuring Windows Vista Power Policy for Maximum Energy Efficiency

Windows Vista allows system manufacturers and IT professionals to configure platforms for maximum energy efficiency by manipulating power policy and other system configuration settings. The default Windows Vista power policy is designed to support a wide breadth of hardware platforms and to balance power savings and energy efficiency with performance, but may be changed to favor performance or power savings and help extend mobile PC battery life.

This section details individual power policies that impact overall system power consumption. For each power policy, the power policy identification globally unique identifier (GUID), its description, and typical power consumption impact are provided. System manufacturers are encouraged to review this information and incorporate power policy changes into their Windows Vista preinstallation images where appropriate. For each power policy change, the power consumption, performance, and user experience impact should be evaluated. Changes to power policy and the impact on system energy efficiency may be evaluated by using the techniques described in the earlier sections of this paper.