Impact of "Idle" Software on Battery Life - Intel

<strong>Impact</strong> <strong>of</strong> “Idle” S<strong>of</strong>twareon Battery LifeManuj Sabharwal, Application EngineerEunice Chang, Engineering ManagerSession ID: EBLS002

Agenda• Importance <strong>of</strong> Battery Life• Role <strong>of</strong> S<strong>of</strong>tware in Battery Life• Case Study: “Idle” S<strong>of</strong>tware Applications Overhead• Tools for Analyzing Power Overhead• Case Studies: Browser, Media Playback, Messaging• Call to Action• Q/A2

3How Critical isLaptop Battery Lifeand the <strong>Impact</strong> <strong>of</strong>S<strong>of</strong>tware?

Top Mobile Computing NeedsSource: Internal Worldwide Market Research, 2010Consumers’ Most Desired Laptop Configuration Improvements(Results when end-used named their top 3)Consume Less Energy/Better BatteryBetter Performance/Improve Speed <strong>of</strong> …More Storage/Bigger Hard DriveBetter Range <strong>of</strong> Wireless Internet …Make it Lighter66%60%45%40%42%Larger Screen/DisplayBetter Appearance/LooksMake it Smaller15%13%22%

S<strong>of</strong>tware Energy-Efficiency• Platform hardware designed with aggressive powermanagement especially when idle (e.g. ACPI, C/Pstates, PCIEASPM, SATA LPM, etc.)• S<strong>of</strong>tware (OS, FW, Drivers) plays important role in platformenergy efficiency• Energy-efficient applications, when idle, should have minimalimpact on platform power consumption– Ideally, application at idle should consume no additionalenergyA single ill-behaving application can thwart all<strong>of</strong> the power management benefits built into system.5

6A Case Study <strong>of</strong>“Idle” S<strong>of</strong>tware Apps

Case StudyThe Measurement System• Intel ® Core2 Duo based Mobile Reference Platform• Core i5 T6600 (Dual Core, 2.2 GHz)• Windows* 7, 64-bit• Intel ® Solid State Drive• Wired LAN• (Power) Fluke* NetDAQThe Measured ApplicationsSelect 25 common applications including:• 4 web browsers• 5 media playback applications• 5 messaging applications• 2 virus scanners7

Study Results5Lower is BetterDeep Dive #2Power Over Baseline (Watts)4321Deep Dive #3Deep Dive #10MessagingPhoto Productivity Plug-ins BrowsersChat Photo Virus Anti-Productivity Plugin Browsers MediaPlayersVirusMedia Players8The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument . Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Battery Life <strong>Impact</strong>10Battery Life (Hours)987650.5W~45Min0.5W~20Min2W~75MinUltraThin, 56WHr, 6W IdleMainstream, 64WHr, 9W Idle40 1 2 3 4 5Additional Power (Watts)Idle power increase has significant battery life impact9

Deep Dive Detail Methodology• Component Power (CPU, Chipset, HDD, LAN/Wifi)• S<strong>of</strong>tware Tools: Hardware, Driver, OperatingSystem, S<strong>of</strong>tware Timers, Memory Utilization, ...– Windows * Powercfg– Battery Life Analyzer– Windows* Performance Analyzer Toolkit(aka “xPerf”)10

Windows* Powercfg• Command-line utility to control the power settings• Inbox with Windows* 7, Leverages ETW• Power Schemes, Hibernate, Standby, Wake Timer / Devices• -ENERGY Option– Platform timers, changes to timers by the application process/dll– Processor Utilization per process– Power Management Settings in Hardware and OS– Needs to run in Administrator Mode, can output raw .etl traces• Attend EBLS003 Mobile Platform Idle Power Optimization –Methodologies and Tools for demo and detailsQuick & Easy to Identify Common Problems11

Battery Life Analyzer• New Tool written by Intel• Simple GUI interface• Identify misbehaving drivers, processes, and hardware thatprevent the platform from entering low power states• Provides power impact estimate based on an algorithm foreach <strong>of</strong>fending piece <strong>of</strong> hardware/s<strong>of</strong>tware• Attend EBLS003 Mobile Platform Idle Power Optimization –Methodologies and Tools for demo and detailsIdentify, Analyze, and Summarize – All in one tool12

Windows* Performance Toolkit - xPerf• Built by Micros<strong>of</strong>t, Available in Windows* 7 SDK• Leverage ETW (Event Tracing for Windows) Technology• Wide range <strong>of</strong> OS performance analysis in a trace file– All processes/threads, User + kernel mode, DPCs and ISRs,Scheduling, Disk and file I/O, Memory, Network, …• Some Issues xPerf can help identify:– Responsiveness issues– Long delays in applications, Slow On/Off transitions– High CPU usage, High Disk usageFeature-rich Tool for Detailed Power Performance Analysis13

14Analysis <strong>of</strong> TwoCase Studies

Investigation #1 – A Browser ApplicationPower Over Baseline (Watts)543212.42.72.6Chipset33%0.8Other7%CPU60%1.4 1.4Deep Dive #12.12.41.62.2 2.34.60MessagingBrowserChat/VoiP PhotoEdit VirusScanProductivity Browser MediaPlayerEdit ScanMedia Players15*The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument . Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

System Idle vs. Browser Idle CPU PackagePowerSystem Idle vs. Browser Idle total CPU Package PowerRuntime graph for core package powerShowing browser power consumptionPower(W)4.003.503.002.502.001.501.000.500.001224364851061271481691902112322532742953163373583794004214424634845055265475685896106316526736947157367577787998208418628839049259469679881009103010511072System Idle PowerBrowser Idle PowerWhy there is an increase in Browser Idlepower from System Idle?Workload Platform Power CPU PowerBrowser Application 12.06 2.63System Idle 10.12 1.4616The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument. Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance

Battery Life Analyzer AnalysisAverage(%)100806040200Increase due toOperational StateactivityAverage C State Residency Change - System idle vs. Browser IdleI/O controlleroperationsBrowser Idlespends less timedeep C statePackage0 C0 - C1 Package0 C2 Package0 C3 Package0 C6Average System idle(%)Average Browser IdleFew Things to Know :• C-states: Collection <strong>of</strong> Idle State.– The deeper the C-States allows to achieve the low powerconsumption, but more time is required for the CPU to “wake up”.• Average C-State Residency is average residency in each C-State overperiod <strong>of</strong> time• Deeper C-state Residency(C6), more power savings (System Idle)• Periodic activity results in more time spent in C0-states (Browser Idle)17

Windows * PowercfgBrowser setting Timer Tick to 1msec18Command line utility giving call stack <strong>of</strong> modulesetting timer tick to 1ms

19Causes: Idle State Power Consumption• Context Switches – Process <strong>of</strong> storing and restoring state <strong>of</strong>a CPU– Increase due to OS requiresto take control <strong>of</strong> the CPU to dostoring and restoring task.– More details on cause <strong>of</strong>high context switches on nextslide• System calls– Increase in periodiccalls/sec due toclock interruptshal.dll: Interrupt Servicing- much higher for application idleImage NameImage NameContext Switchesat system idleSystem IdleCalls/secContext SwitchesBrowserPlatform Activity 53 550BrowserCalls/sechal.dll 64.1 999.9ataport.SYS 3.0 1.0ACPI.sys 0.4 0.4USBPORT.SYS 3.0 2.9

Windows* Performance Toolkit - xPerfAnalysis for BrowserxPerf shows frequent CPU activities for browser app due tomany periodic activitiesProcess Timer Interval RemarksBrowser.exe 150msec GetMessageW21msec20msecSendMessageSleepBrowser Service 4 msec FlushFileBuffers20 msec WaitForSingleObject()Alternatives?-Optimized Sleep version to give better performance and power.-Use user land locks if possible instead <strong>of</strong> WaitForSingleObject-Use unbuffered I/O instead <strong>of</strong> calling FlushFileBuffersCan periodic intervals prolonged and their activities coalesced ?20

Effect Of Timer-Tick On Browser ApplicationTimer-Tick Value Platform Power CPU PackagePower1msec (Default app Idle) 12.06 2.6310msec 11.04 1.95System Idle 10.12 1.46• Timer-Tick change saves 1.02 Watts <strong>of</strong> total power at platform level.Out <strong>of</strong> Box, this browser uses 60 extra minutesbattery life at “idle” on a standard 64Whr battery21The Power comparisons shown is based on the Intel Core i5- T6600 processor using NetDAQ® Fluke ® instrument. Results have been estimated based on internal Intel analysis andare provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Analysis• Application appears to be setting a timer interval <strong>of</strong> 1ms• The total average CPU utilization < 0.2%• Many periodic activity with small intervals causing CPU wakeup bypolling driversRecommendations:• Don’t make the timer tick change request until it’s necessary• Use event driven code rather than polling CPU• Review the necessity <strong>of</strong> periodic activity and minimize• Use un-buffered I/O instead <strong>of</strong> calling FlushFileBuffers22

Investigation #2 – A Playback ApplicationPower Over Baseline (Watts)543212.42.72.60.81.4 1.4Other22%Chipset39%2.12.4CPU39%1.6Deep Dive #22.2 2.34.60MessagingBrowserChat/VoiP PhotoEdit VirusScanProductivity Browser MediaPlayerEdit ScanMedia Players*The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® . Results have been estimated based on internal Intel analysis and are23provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Battery Life Analyzer Analysis100806040200C-State Package Residency- System Idle vs. Playback IdleApp preventing CPU to alow power sleep statePackage0 C0 - C1 Package0 C2 Package0 C6System IdlePlayback Idle• CPU can enter different idle power states• Lower power states use less power but the CPU will take longer totransition to a higher power state.– Example – If CPU is in C6 it will use less power than in C3 but it willtake longer to come out <strong>of</strong> idle than from C3Top Cause for CPU to wakeup can be analyzed using xPerfand Battery Life Analyzer24

Windows* PowercfgPlayback Setting Timer Tick to 1.25msec25

Battery Life Analyzer for PlaybackCause 1: Increase in Context switchesfrom idle due to Timer tick andI/O ActivityImage NameContextSwitches atsystem idleContextSwitchesPlaybackPlatformActivity 53 689Causing graphics frame bufferupdate frequently which iscalled by playback application.Cause 2: Increase in periodiccalls/sec due to clock interruptsImage NameSystem IdleCalls/secPlaybackCalls/sechal.dll 64.1 800dxgkrnl.sys 1 210USBPORT.SYS 10 18ataport.SYS 3.0 4.2At idle reduce the periodic graphics activitysuch as animation26

Windows* Performance Toolkit – xPerf AnalysisProcess/ Driver Timer Interval RemarksPlayback.exe 10msec Sleep1()33msec31msecSleep2()Sleep3()xGUISystem 50msec Sleep()D3D9System 1msec Graphics driver call from Application• Avoid calling SetThreadExecutionState(ES_DISPLAY_REQUIRED) when thereis no playback activity, i.e. at idleTimer-Tick Value Platform Power CPU PackagePower1msec(App Default)System Idle(15.6msec)14.06 4.4910.12 1.4627Out <strong>of</strong> Box, this playback uses 112 extra minutesbattery life at “idle” on a standard batteryThe Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument. Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Analysis• Application appears to be setting a timer interval <strong>of</strong> 1.25ms• Application having frequent periodic activity by calling Sleep API• Multiple threads are looping at Sleep() independently with variousinterval. All combined, this process wakes up the processor veryfrequently• Application doing a lot <strong>of</strong> memory copy between CPU & GPU memoryRecommendations:• Do not try to change timer tick from 15.6 ms default when application is atidle.• Use event driven API calls than Polling.• Instead <strong>of</strong> running multiple thread periodically (Sleep() API ) run onethread at the lowest frequency possible, and if necessary, signal to otherthread.• Do not render to the display when it is <strong>of</strong>f• The system display might be <strong>of</strong>f for power savings. Your applicationshould not perform unnecessary graphics rendering when the displayis <strong>of</strong>f because this wastes system resources and power.28

Study Results SummaryPower Over Baseline (Watts)54321Lower is BetterMessenger having low powerconsumption when IdleBrowser having low powerconsumption when IdlePlayback having low powerconsumption when Idle0MessagingPhoto Productivity Plug-in BrowsersChat Photo Virus Anti-Productivity Plugin Browsers MediaPlayersVirusMedia Players29Why can’t Application behave like good ones in theirown categories?*The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument. Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Summary• Battery Life is an important attribute <strong>of</strong> Laptopsystems• Case Studies shows power impact at applicationidle on battery life.• Changes can be made that do not affect s<strong>of</strong>twareperformance and reduce the impact. E.g.– Avoid use <strong>of</strong> high-resolution periodic timer– Do not render to the display when it is <strong>of</strong>f– Avoid polling and spinning in tight loops– Use timer coalescing• Reduce the background application serviceswhen idle.30

Call to Action• Evaluate the idle impact <strong>of</strong> your s<strong>of</strong>twareapplications – pay specific attention to timer-tickinterrupts• Ask your suppliers/Vendors to do the same andaware them about power awareness.• Blog, Ask questions in forums, raise awarenessabout application that consume too much power• Promote Green S<strong>of</strong>tware• Join Extended Battery Life Working Group(www.eblwg.org)31

33IDF 2010 Extended Battery Life SessionsSession # Title Time RoomEBLS001Interconnect Bus Extensions forEnergy-Efficient Platform10:15 AM 2003EBLS002<strong>Impact</strong> <strong>of</strong> “Idle” S<strong>of</strong>tware on BatteryLife11:15 AM 2003EBLS003EBLP001Mobile Platform Idle PowerOptimization – Methodologies andToolsEmbedded Display Port: NextGeneration <strong>of</strong> Digital DisplayInterface for Embedded Application1:05 PM 20032:05 PM 2003

Legal Disclaimer• INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL® PRODUCTS.NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUALPROPETY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL’S TERMSAND CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITYWHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TOSALE AND/OR USE OF INTEL ® PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TOFITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT,COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT.• Intel may make changes to specifications and product descriptions at any time, without notice.• All products, dates, and figures specified are preliminary based on current expectations, and aresubject to change without notice.• Intel, processors, chipsets, and desktop boards may contain design defects or errors known aserrata, which may cause the product to deviate from published specifications. Currentcharacterized errata are available on request.• Performance tests and ratings are measured using specific computer systems and/orcomponents and reflect the approximate performance <strong>of</strong> Intel products as measured by thosetests. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actualperformance.• Intel Core i5 Mobile Processor and the Intel logo are trademarks <strong>of</strong> Intel Corporation in theUnited States and other countries.• *Other names and brands may be claimed as the property <strong>of</strong> others.• Copyright ©2010 Intel Corporation.34

Risk FactorsThe above statements and any others in this document that refer to plans and expectations for the second quarter, the year and thefuture are forward-looking statements that involve a number <strong>of</strong> risks and uncertainties. Many factors could affect Intel’s actualresults, and variances from Intel’s current expectations regarding such factors could cause actual results to differ materially fromthose expressed in these forward-looking statements. Intel presently considers the following to be the important factors that couldcause actual results to differ materially from the corporation’s expectations. Demand could be different from Intel's expectations dueto factors including changes in business and economic conditions; customer acceptance <strong>of</strong> Intel’s and competitors’ products; changesin customer order patterns including order cancellations; and changes in the level <strong>of</strong> inventory at customers. Intel operates inintensely competitive industries that are characterized by a high percentage <strong>of</strong> costs that are fixed or difficult to reduce in the shortterm and product demand that is highly variable and difficult to forecast. Additionally, Intel is in the process <strong>of</strong> transitioning to itsnext generation <strong>of</strong> products on 32nm process technology, and there could be execution issues associated with these changes,including product defects and errata along with lower than anticipated manufacturing yields. Revenue and the gross marginpercentage are affected by the timing <strong>of</strong> new Intel product introductions and the demand for and market acceptance <strong>of</strong> Intel'sproducts; actions taken by Intel's competitors, including product <strong>of</strong>ferings and introductions, marketing programs and pricingpressures and Intel’s response to such actions; defects or disruptions in the supply <strong>of</strong> materials or resources; and Intel’s ability torespond quickly to technological developments and to incorporate new features into its products. The gross margin percentage couldvary significantly from expectations based on changes in revenue levels; product mix and pricing; start-up costs, including costsassociated with the new 32nm process technology; variations in inventory valuation, including variations related to the timing <strong>of</strong>qualifying products for sale; excess or obsolete inventory; manufacturing yields; changes in unit costs; impairments <strong>of</strong> long-livedassets, including manufacturing, assembly/test and intangible assets; the timing and execution <strong>of</strong> the manufacturing ramp andassociated costs; and capacity utilization. Expenses, particularly certain marketing and compensation expenses, as well asrestructuring and asset impairment charges, vary depending on the level <strong>of</strong> demand for Intel's products and the level <strong>of</strong> revenue andpr<strong>of</strong>its. The majority <strong>of</strong> our non-marketable equity investment portfolio balance is concentrated in the flash memory market segment,and declines in this market segment or changes in management’s plans with respect to our investment in this market segment couldresult in significant impairment charges, impacting restructuring charges as well as gains/losses on equity investments and interestand other. Intel's results could be impacted by adverse economic, social, political and physical/infrastructure conditions in countrieswhere Intel, its customers or its suppliers operate, including military conflict and other security risks, natural disasters, infrastructuredisruptions, health concerns and fluctuations in currency exchange rates. Intel’s results could be affected by the timing <strong>of</strong> closing <strong>of</strong>acquisitions and divestitures. Intel's results could be affected by adverse effects associated with product defects and errata(deviations from published specifications), and by litigation or regulatory matters involving intellectual property, stockholder,consumer, antitrust and other issues, such as the litigation and regulatory matters described in Intel's SEC reports. An unfavorableruling could include monetary damages or an injunction prohibiting us from manufacturing or selling one or more products,precluding particular business practices, impacting our ability to design our products, or requiring other remedies such as compulsorylicensing <strong>of</strong> intellectual property. A detailed discussion <strong>of</strong> these and other factors that could affect Intel’s results is included in Intel’sSEC filings, including the report on Form 10-Q for the quarter ended March 27, 2010.Rev. 5/7/1035

36Backup

Investigation #3 – A Messenger ApplicationPower Over Baseline (Watts)54321Deep Dive #32.42.72.6Other29%Chipset26%0.8CPU45%1.4 1.42.12.41.62.2 2.34.60MessagingBrowserChat/VoiP PhotoEdit VirusScanProductivity Browser MediaPlayerEdit ScanMedia Players*The Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument. Results have been estimated based on internal Intel analysis37and are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance/Power

Battery Life Analyzer AnalysisAverage(%)100.0080.0060.0040.0020.000.00C-State Package Residency- System Idle vs. Messenger IdleMore time spent in deeper C-Statecan result in better powerPackage0 C0 - C1 Package0 C2 Package0 C6Average System Idle (%) Average Messenger Idle (%)Increase in C0-State Package Residency is same as in previous Case studiesCause: Context switchesfrom idle due to Timer tick andI/O ActivityImage NameContextSwitches atsystem idleContextSwitchesMessengerPlatform Activity 53 1138Cause 2: Increase in periodiccalls/sec due to clock interruptsImage NameSystem IdleCalls/secMessengerCalls/sechal.dll 64.1 1000USBPORT.SYS 10 12ataport.SYS 3.0 4.238A break event brings the idle CPU to C0 State

Windows * PowercfgMessenger Setting Timer-Tick to 1msec39

Windows* Performance Toolkit - xPerfAnalysis for MessengerProcess Timer Interval RemarksMessenger.exe 165msec GetMessage2msec2msec20msecWaitForSingleObjecttimeSetEventRegQueryValue40Alternative?:•Use Event Driven code - RegNotifyChangeKeyValue() instead <strong>of</strong> RegQueryValue•Use <strong>of</strong> coalescing timer API.Timer-Tick Value Change EffectTimer-Tick Value Platform Power CPU Package Power Memory1msec(ApplicationDefault) 11.95 2.48 1.1210msec10.71 1.87 0.31System Idle(15.6msec)10.12 1.46 0.08Out <strong>of</strong> Box, this Messenger uses 58 extra minutesbattery life at “idle” on a standard batteryThe Power comparisons shown is based on the Intel® Core i5 Mobile Processor using NetDAQ® Fluke ® instrument . Results have been estimated based on internal Intel analysisand are provided for informational purposes only. Any difference in system hardware or s<strong>of</strong>tware design or configuration may affect actual performance

Analysis• Application appears to be setting a timer interval <strong>of</strong> 1ms• Application having frequent periodic activity – WaitForSingleObject,GetMessageW, TimerSetEvent.• I/O Reads/Writes to the registry increases 2x time on average compared to systemidleRecommendations:• Do not change Time Tick value from system default.• Minimize the use <strong>of</strong> polling. If possible, use an interrupt instead.• If the driver must poll, consider implementing timers by using the newtimer coalescing techniques which reduces the CPU wake frequency• Periodic registry activity must also be avoided• Registry activity eventually results in disk activity41

Timer Coalescing• Platform energy efficiency can beimproved by extending idle periods– New timer coalescing API enables callersto specify a tolerance for due time– Enables the kernel to expire multiple timers at the sametime• Extensions should integrate with Windows* 7API/DDI