back to article Chrome browser has been DRAINING PC batteries for YEARS

If you've been wondering why the battery life on your Windows laptop or tablet seems so lousy, your Chrome web browser might be to blame – and it may have been sapping your system's juice for years. A documented bug in the source code for the Chromium open source project seems to account for the mysterious power drain that …

  1. Anonymous Coward
    Headmaster

    PC battery? You mean the coin battery backing up the CMOS? Or do you consider the desktop PC dead now and PC means laptop?

    1. This post has been deleted by its author

      1. Anonymous Coward
        Anonymous Coward

        Re: I have two 12V cells

        I bet you don't, I bet they're batteries of several cells.

    2. big_D Silver badge

      Good, then in your case, artificially driving up the electricity bill...

    3. Anonymous Coward
      Anonymous Coward

      My UPS battery is my "PC's battery". The bug would affect that too.

      1. Gis Bun

        Not really, unless you have a power failure.

  2. Anonymous Coward
    Anonymous Coward

    I, personally, am not surprised

    Chrome had sucked for years, regardless of the fanboi-level user base. I have complained about it's miserable RAM cache management, even worse than Firefox's (no mean feat), and known for years

    http://www.google.com/m?q=chrome%20ram%20usage

    as well as the dumbed-down user interface, which is just...dumb.

    So now Google, after years of knowing, finally fixes a battery drain issue, a rather important issue in today's mobile world.

    Thanks Google. Your browser STILL sucks.

    1. Breen Whitman

      Re: I, personally, am not surprised

      "even worse than Firefox"

      Don't be so preposterous. It lowers one's credibility.

      1. Dan 55 Silver badge

        Re: I, personally, am not surprised

        Firefox is well-behaved these days regarding memory. Chrome most certainly isn't.

        http://www.ghacks.net/2014/01/02/chrome-34-firefox-29-internet-explorer-11-memory-use-2014/

        1. phil dude
          Meh

          Re: I, personally, am not surprised

          and yet, per process tabs are not standard....

          I have had far fewer crashes with chrome. I keep firefox because I need some plugins to work...

          P.

          1. phil dude
            WTF?

            Re: I, personally, am not surprised

            I run chrome and chromium (the FOSS version) with many, many 100's of tabs, because if one dies it does not take the ship down with it.

            Firefox has improved greatly, but until a broken piece of software does not take down the browser, the only reason I use it is for some plugins that chrome/chromium do not support.

            I am well aware they are trying to implement it (electrolysis) but until it is standard and stable, it will continue to be a problem for Firefox stability.

            P.

        2. Anonymous Coward
          Anonymous Coward

          Re: I, personally, am not surprised

          > Firefox is well-behaved these days regarding memory. Chrome most certainly isn't.

          I must be running a different version to you. Currently I've 9 tabs open and it's consuming more than 1.8GB of RAM. It's barely come down since I closed 10-12 or so others.

          I know if I restart it that'll drop down to a few hundred meg.

          Been like this as long as I can remember. I always have to restart Firefox every few days to allow it to clean itself up.

          1. Tom 38

            Re: I, personally, am not surprised

            Some people only run Firefox for a few hours or so and do not realize how awful it still is at leaking memory when used for days or weeks. Use it without restarting for a week or more, close down all tabs, and it will still be using gigabytes of RAM.

          2. Muscleguy
            Devil

            Re: I, personally, am not surprised

            Me too, I usually crash out of it as this means I have more chance of being able to choose which windows and tabs to reopen before doing so.

            HOWEVER, today I upgraded from 30.0 to 31.0 and on reopening every single window had an additional tab from a company called abine.com offering do not track me add-ons. Closing them in about a third of the windows caused new ones to spawn. Now when tabs are reloaded some of them cause more respawnings. I have alerted Mozilla and requested they remove this crap they have installed.

            Oh and now, just like this window, the bottom add-on bar is not there.

            At this rate I will be forced back to Safari, oh the horror!

          3. KayKay
            Meh

            Re: I, personally, am not surprised

            about:config

            right click, >> new >>> Boolean

            config.trim_on_minimize

            true

            open all windows and minimize them, and watch the RAM usage plummet. If you don't leave open web mail tabs, and kill (through Task Manager) the plugin container when not watching videos, you can go up to 100 tabs in bursts, and not have to reload the browser for 9 or 10 days at a time. Running AdBlock also helps.

            EDIT: Oh and don't used Firefox. If you have Windows you have the choice of Pale Moon and Waterfox, FF variants reset to having normal behavoiur.

            EDIT: SVG support in those, too.

            Google is your friend. Well not when they're designing Chrome, but their search engine is still okay.

        3. Wyrdness

          Re: I, personally, am not surprised

          "Firefox is well-behaved these days regarding memory"

          I came in to work this morning and found my (Linux) PC almost completely unresponsive.

          top showed that Firefox was using 95% of the 12GB ram and, including swap space, a total of 17GB.

          So about as well-behaved as a ravenous sabre-toothed tiger.

    2. Turtle

      Re: I, personally, am not surprised

      "So now Google, after years of knowing, finally fixes a battery drain issue, a rather important issue in today's mobile world."

      One of us has misunderstood the story. From what I read, the bug has not been fixed; it has been assigned to someone to be fixed. So, depending on the person to whom it has been assigned, and the state and severity of their substance abuse problems, it could be another few years before a fix is actually implemented and made available to end-users.

    3. h3

      Re: I, personally, am not surprised

      Google is awful when it comes to power management. (Putting the stuff from Latitude into Google Play Services has made what was a 4/5 day battery life into about a day and a bit.)

      They use the same sort of update mechanisms as malware as well.

    4. This post has been deleted by its author

  3. Notas Badoff

    Sometimes, nobody inside cares what *you* think

    I've seen this before. One or another person on an open source dev team decides that something is just "not a problem" and it doesn't matter how many outside people say "WTF?!" No arguments will help to move the issue. In this article's case only shaming the entire team worked.

    In my favorite illustration is the "Firefox won't do SVG fonts" issue. Opened in 2002, after ~150 comments it was closed in 2013 with "Oh, poo, implementation of SVG embedded in OpenType fonts will fix your problem." An experimental proposed feature that is still not standardized or accepted by other browsers over a year later.

    It didn't even matter when an outside company offered to fund the development of an implementation with their own devs, FF devs simply said "Oh, but SVG fonts are just so ucky!" "Why, you can't even do font hinting with SVG, can you?" So obviously no implementation was ever so much better than a flawed implementation, right?

    And the pleas kept coming in, with increasingly exasperated replies of "but you just don't understand our (FF devs) position of 'No'. How rude of you!"

    So over 12 years later FF still doesn't have anything to offer and will not be swayed!. A big enough closed-minded open source team will quite well replicate the worst parts of the commercial experience.

    1. Steve Davies 3 Silver badge

      Re: Sometimes, nobody inside cares what *you* think

      If you are really pissed off by the refusal of Mozilla to implement SVG fonts then the solution is simple...

      Fork it and create your own browser with the feature included. If you get enough support then it will become popular and hey, who knows they might include it.

      No? Well don't complain about the lack of functionality then.

      1. Anonymous Coward
        Anonymous Coward

        Re: Sometimes, nobody inside cares what *you* think

        Are you saying you can't complain about something unless you're willing to build it yourself?

        What a strange concept. Not everyone has the capability, time, or inclination to fork a browser. (Easy though it might be).

        Imagine if that line of thinking was applied to everything you use.

        'This TV is great apart from the fact that it's filling up my entire room, ideally I'd like a flat TV, but my TV manufacturer says no one needs a flat TV... I know I'll build my own...oh.. I've got a life to be getting on with. I'll just keep quiet and learn to love my room eating monster..'

        Innovation must work around what the users want!

        1. DN4

          Re: Sometimes, nobody inside cares what *you* think

          > Are you saying you can't complain about something unless you're willing to build it yourself?

          You can complain as much as you want. But you have the full right to be ignored if you are *just* complaining.

          As an open source developer (not paid by any rich foundation, mind you), I often must say: I do not want to invest my time into this. There are more important/interesting things to work on and this one is weird/uninteresting/marginal/tangential to the program function/... But patches are welcome. This is how open source works.

          Have you tried your complaining tactics with proprietary software? How it went?

      2. Destroy All Monsters Silver badge
        Mushroom

        Re: Sometimes, nobody inside cares what *you* think

        No? Well don't complain about the lack of functionality then.

        Listen up, a*****e!

        One of parties involved sits on a stash of a few million bucks of "foundation money" that it is rubbing its body parts all over.

        The other does not and probably has a day job.

        Care to speak up again?

      3. RealFred

        Re: Sometimes, nobody inside cares what *you* think

        And this si the typical arrogance of Open Source developers. They want people to use their products, but don't want to give users what they want. I've just dropped using Firefox and Chrome because they suck so much. If thats the standard of Open Source applications, then they really aren't an alternative.

    2. Anonymous Coward
      Anonymous Coward

      Re: Sometimes, nobody inside cares what *you* think

      https://bugzilla.mozilla.org/show_bug.cgi?id=150025 has been open for 12 years. I guess the fix to stop http://localhost redirecting to http://www.localhost.com is just TOO HARD. Or perhaps not a priority. I mean, there's a new interface paradigm to code every two months!

      1. Anonymous Coward
        Anonymous Coward

        Re: Sometimes, nobody inside cares what *you* think

        That isn't a firefox bug. By default, firefox will try to resolve "localhost" and if if fails it will add the www and .com. It does this for any single word you type in. This issue occurs because your PC is not resolving localhost to 127.0.0.1. Fix your PC and that behaviour will stop.

        1. Anonymous Coward
          Anonymous Coward

          Re: Sometimes, nobody inside cares what *you* think

          "That isn't a firefox bug. By default, firefox will try to resolve "localhost"

          Yes it IS a bug. http://localhost is ALWAYS 127.0.0.1 There's no resolving to do at all.

          1. Anonymous Coward
            Anonymous Coward

            Re: http://localhost is ALWAYS 127.0.0.1

            So why do we have to put it in /etc/hosts (or equivalent) and why do various applications look for it there, if it never changes?

          2. Mike Pellatt
            Facepalm

            Re: Sometimes, nobody inside cares what *you* think

            How is "localhost" ALWAYS turned into 127.0.0.1 ?? By calling the, errr, resolver library. Which will obey whatever rules are in its configuration to turn the string "localhost" into an IPv4 or IPv6 address.

            This is known as resolving.

            Or did I miss something ironic in your comment ?

          3. Anonymous Coward
            Anonymous Coward

            Re: Sometimes, nobody inside cares what *you* think

            > Yes it IS a bug. http://localhost is ALWAYS 127.0.0.1 There's no resolving to do at all.

            The entire 127.0.0.0/8 address block has been reserved for the loopback interface. Each and every one the the 16+ million IP addresses will be looped back. Any one of them can be defined and used as localhost (see RFC 5735 Special Use IPv4 Addresses).

            Most OSes (this includes MS) have a hosts file that contains the lookup for localhost. It should always be resolved.

            So no it ISNT a bug.

            1. Anonymous Coward
              Anonymous Coward

              Re: Sometimes, nobody inside cares what *you* think

              Also, localhost in ipv6 doesn't resolve to 127.0.0.1

              This should be handled by the o/s, not by the browser.

          4. Jeff Green

            Re: Sometimes, nobody inside cares what *you* think

            localhost doesn't have to be 127.0.0.1, it can be any address, although to fit with the standards that should be within the 127.0.0.1/8 range. On my machine it is 127.0.1.1 because I am too lazy to rewrite a very badly written hack I created years ago that used 127.0.0.1 ...

      2. SlimPickens

        Re: Sometimes, nobody inside cares what *you* think

        Have you tried recently? No matter what port I add to the end of localhost it never tries to go out to localhost.com...

    3. pip25
      Unhappy

      Re: Sometimes, nobody inside cares what *you* think

      I feel this is a more general issue and seems to be part of a trend I noticed with projects which are supposedly open to contributions.

      These projects have several types of users in terms of the available time they commit to it, but inevitably, there will be at least a few "fanatics" who do not mind the project taking up a huge amount of their time, quickly become leaders and organizers, and are responsible for much of the project's momentum.

      Unfortunately, it seems like they are also the very same people who will be responsible for the project's eventual downfall. Why? Simply because, after all the time and work they've put into it, they now feel that the project (or at least a good chunk of it) belongs to them. This is in fact a completely understandable emotion given the circumstances, but also very destructive, since instead of trying to inspect a proposal based on its own merits, they will became increasingly prone to dismissing something which doesn't harmonize with their current "vision" of the project out of hand. This not only discourages other, less "fanatic" contributors, but can also lead to the degradation of the project's overall quality, since, let's face it, even these very smart and talented people can be wrong, but with their status and mindset they can become largely immune to critique.

      I've seen this attitude in so many places that it's kind of scary. Firefox and Chrome are an example, but then there's the whole recent Gnome debacle, too. And it's not just about software projects: I feel Wikipedia suffers from the very same problem, with its editors becoming increasingly elitist and seeing new people as more of a hindrance than an asset. Heck, I'm an admin of a smaller wiki too and I've seen the very same effect on myself.

      Without wanting to sound overly dramatic, unless we find some kind of solution to this problem, this can eventually have a devastating effect on the open-source community as a whole. :(

  4. Anonymous Coward
    Anonymous Coward

    Of Google software and battery drain

    While we're on this topic:

    Google, when are you going to fix the shameless power-sucking pig also known as Google Play on Android devices? That useless piece of bloatware is responsible for battery drain on almost all JellyBean and KitKat devices.

    Perhaps you could start by explaining what Google Play actually does that it needs so much power. Does it calculate PI to the 1024th decimal in an infinite loop?

    Or this is not not considered a problem in Mountain View?

    1. Anonymous Coward
      Anonymous Coward

      Re: Of Google software and battery drain

      >Perhaps you could start by explaining what Google Play

      Perhaps you could explain why you *think* it's google play that is using all that power..

      1. Anonymous Coward
        Anonymous Coward

        Re: Of Google software and battery drain

        > Perhaps you could explain why you *think* it's google play that is using all that power..

        I don't *think* it is Google Play, I *know* it is Google Play, and so does Sony, amongst others. Google Play has been outed as the major pig that it is officially, by Sony Mobile:

        http://www.xperiablog.net/2014/06/05/sony-acknowledges-xperia-z-kitkat-battery-woes-provides-workaround/

        The "workaround" is to disable Google Play Services.

        1. Anonymous Coward
          Anonymous Coward

          Re: Of Google software and battery drain

          That appears to be a problem specific to Sony devices. Not sure how that is Google's fault. It's very possible that something in the Play Services triggers something in Sony devices that cause them to fully wake and that causes extra battery usage. The thing is the play services are used by other apps so it may very well be that one of Sony's shovel-ware apps is what is causing play services to be active... It's sort of like blaming the milk for running out when you open the bottom of the carton.

          1. david bates

            Re: Of Google software and battery drain

            It also impacted my 2012 Nexus 7 on KitKat and on Cyanogenmod 11.... Even with wireless and such disabled it would drain the battery in a couple of days on standby. It would routinely use at much as the screen in normal use. I have no idea what it was doing as it has no connection to the outside world.

            Disabling it has fixed the problem, so yes, its play services, no, its not a Sony problem...

            1. Jamie Jones Silver badge

              Re: Of Google software and battery drain

              I find that Googles own Android apps are the worst offenders.

              Only today I had to kill G+ because it was sucking CPU and I/O. Maps was also running - I've used neither in months.

              Android is my tablet OS of choice, but I'd be pretty pissed off at it if I didn't have the ability to keep it under control.

            2. Blitterbug
              Thumb Up

              Re: Disabling it has fixed the problem, so yes, its play services...

              Indeed it is. Can confirm same on my 2012 Nexus 7, 2013 (I think) Galaxy Tab 3 10", and my new Xperia Z-1 tablet. I have to go in and disable all the Play apps (which I never use, so no issue there) and G+ just for luck. Battery life quickly returns to something respectable.

          2. Anonymous Coward
            Anonymous Coward

            Re: Of Google software and battery drain

            > Not sure how that is Google's fault.

            The Bullshit comes out swinging. You must work at Google.

    2. dotdavid

      Re: Of Google software and battery drain

      The OP means Google Play Services, which is pretty much a kitchen sink of things that Google thinks Android developers might find useful. Some of those things impact battery life, most probably the location-related stuff (formerly known as Latitude).

      I found that the Google Play Services process was keeping a wakelock for ages, thus the phone was never "sleeping" and the battery was draining quickly. I disabled the permission in CyanogenMod AppOps (unfortunately not available to any stock ROM, although you can root and install it) and sure enough the problem went away.

      There is speculation online that this wakelock problem is actually caused by some other apps that are *using* Google Play Services, unfortunately AFAIK it's impossible to tell - one downside of this library approach to system functionality I guess. Would be helpful if there was some kind of statistics monitor available for Google Play Services though.

  5. Anonymous Coward
    Anonymous Coward

    Sounds like an issue with how Windows uses the platform timer opposed to Chrome causing the battery drain.

    Why is something that can cause issues like this available to user applications in the first place?

    1. Anonymous Coward
      Anonymous Coward

      My thoughts exactly.

    2. silent_count

      It's hardly Windows' fault if people choose to use a battery draining program. The user has presumably chosen to install a program which behaves in this manner.

      I'm no MS appologist but, in this case, it really isn't their fault. If they don't allow programs to change the timer resolution, someone's 'cure for cancer' program will depend on being able to change the timer resolution, and then MS are autocratic pricks who won't let people use the hardware which they've paid good money for.

      If they do allow it, someone (hi there, Google) will use it for something stupid and MS gets the blame for allowing them to.

    3. Charles Manning

      I doubt the sleep issue too

      I really doubt that most CPUs are really sleeping the full 15mS anyway.

      Every time there is an interrupt (mouse wiggle, network packet, end of video frame,...) the CPU has to wake up to process it.

      If the timer is just being woken up to service Chrome, then it should be going to sleep again fast.

      Sure there are some corner conditions (ie. no network, no mouse activity) when this bug might have been causiung a few % change in power usage, but that would be the exception and you can understand why the developers would give this a low priority.

    4. This post has been deleted by its author

    5. eldakka

      Err, because some apps that you install might NEED to check state more often than the default? near-real time systems might want to wake every 4 or 5 ms, or even as google chrome thinks it needs to, every 1ms.

      Unless you've loaded a plugin into chrome that for some reason needs frequent wake-ups, then there's no reason for a browser to want to wake up more frequently than the defauly.

  6. aberglas

    But isn't it event driven?!

    Does anybody know what this article is actually talking about? An operating system is normally event driven, ultimately from interrupts that come from external sources such as keyboards, network and timers. There should be no polling.

    Certainly if Chrome sets up a polling loop with 1ms resolution then windows should probably obey and that will kill the battery. But why on earth would Chrome poll for anything?!

    1. Anonymous Coward
      Anonymous Coward

      Re: But isn't it event driven?!

      >An operating system is normally event driven,

      Timers generate events.

      >ultimately from interrupts that come from external sources such as keyboards,

      >network and timers. There should be no polling.

      How do you do pre-emptive multitasking if external interrupts are the only way to jump out of the running user task and re-enter the kernel?

    2. Anonymous Coward
      Anonymous Coward

      Re: But isn't it event driven?!

      >Chrome sets up a polling loop with 1ms

      I think I misunderstood your post because you have misunderstood what the issue is.

      It seems that Chrome plays with the platform timer. That's not a "polling loop".

      If a user process playing with that is an issue it shouldn't be available to user processes unless they run as a super user.

  7. chivo243 Silver badge

    It does affect OSX

    I was just trouble shooting a friend of the missus MBP, she had fast battery drain. Both Chrome and Skype were set to launch at login. When clicking on the battery icon the menu said: Applications using significant power: Google Chrome

    With Chrome running on that MBP, the charger could not charge the battery and run Chrome at the same time. Once I quit Chrome the battery resumed charging normally.

    1. Anonymous Coward
      Anonymous Coward

      Re: It does affect OSX

      Nice bit of clueless user level troubleshooting. Just because chrome on your machine is using significant battery does not mean that it is due to the same bug. It might be a different bug or it might be your setup.

      1. Anonymous Coward
        Anonymous Coward

        Re: It does affect OSX

        http://en.wikipedia.org/wiki/Correlation_does_not_imply_causation

        1. VeganVegan

          Re: It does affect OSX

          Correlation != causation

          Obligatory XKCD reference: http://xkcd.com/552/

          1. bazza Silver badge

            Re: It does affect OSX

            "Correlation != causation"

            Indeed not, though given that the code base is (presumably) quite similar for the OS X variant there's a good chance that it is the same issue.

            1. Dan 55 Silver badge

              Re: It does affect OSX

              There's no way Chrome on Mac has the same problem as Windows on Mavericks (it's Mavericks because only Mavericks has the battery usage hall of shame in the battery menu) because Mavericks also has a new way of conserving battery power that specifically groups CPU wake-up timers together. It would refuse to wake up the CPU every ms just for Chrome, instead it would bunch Chrome's wake-up timers with other apps at the cost of less accuracy.

              It's a different problem, but the same root cause (crappily programmed).

              1. ThomH

                Re: It does affect OSX @Dan 55

                Apple's implementation of timer eliding relies on the application switching from saying "I want a timer that fires every X milliseconds" to "I want ... plus or minus Y milliseconds". Applications that don't specify the tolerance continue to receive the old behaviour. Knowing Apple that's likely only a transitional move but is in effect as of v10.9.

                Incidentally, anecdotal observations: one uses significant battery power by waking often regardless of the amount of work done. An application I wrote woke regularly 50 times a second but used barely 3% CPU time — it was an emulator, doing the most obvious thing. It got into the hall of shame. Switching to an adaptive timer, only when wakes drop below about 10 times a second does it stop being named as a problem. Of course, that's empirical without being particularly scientific and I've quite possibly made a dumb mistake elsewhere, etc, etc.

                1. Dan 55 Silver badge

                  Re: It does affect OSX @Dan 55

                  Re-reading the review on Ars, apparently Cocoa and Mavericks decide if an old app should behave as before or with the new behaviour, although you can disable the new behaviour in the Get Info box.

                  http://arstechnica.com/apple/2013/10/os-x-10-9/14/

          2. Anonymous Coward
            Anonymous Coward

            Correlation != causation

            True but... If there actually *is* causation, the first clue is often correlation. 'Correlation is not causation' was the refuge of the tobacco companies for decades. This 'Correlation != causation' advice has, through overuse, swung into negative value territory. It presently supports the situation that nothing can ever be proven on the 'net. Woof.

            (The above are general comments, not necessarily applicable to the present discussion.)

            1. Anonymous Coward
              Anonymous Coward

              Re: Correlation != causation

              "True but... If there actually *is* causation, the first clue is often correlation."

              It doesn't surprise me if my web browser uses most of my computer's CPU power. Most of the time I'm reading/writing emails with my computer, editing text documents, etc. which requires basically zero CPU power, whereas having web pages up in the background is a constant drain on CPU power since most web pages these days are (unfortunately) constantly running some JavaScript and/or displaying ads, etc. So it would actually be a pretty big surprise to me if Chrome DIDN'T use most of my CPU power.

    2. Anonymous Coward
      Anonymous Coward

      Re: It does affect OSX

      >>I was just trouble shooting a friend of the missus MBP, she had fast battery drain. Both Chrome and Skype were set to launch at login. When clicking on the battery icon the menu said: Applications using significant power: Google Chrome

      Well, Chrome might have a timer bug or whatever, but its main thing is rendering web pages, and some of those web pages are enormously complex and suck down huge CPU power. Depending on which web pages you have open, Chrome could be using all of your CPU power or almost none of it. Just because it's showing up in some power drain list doesn't point to why.

  8. Christian Berger

    We should fix the underlying problem

    Browsers are just so complex, they have such huge code bases nobody can properly understand them any more. Maybe we should go on, and split up a successor to HTML into 2 separate standards. One for "documents" and the other one for applications. Both would need to have to be as simple as possible to get rid of legacy bad design decisions like being able to execute code from other webservers than the one you load your HTML from.

    Maybe we could then write a client for that new protocol in a browser, so there would be a trivial way to transition over to it.

    1. ld0614

      Re: We should fix the underlying problem

      How about creating a native way for applications to communicate with the operating system, this would make them much faster! We could then store documents on the operating system and allow the applications to access them.

      wait a minute...

      I think someone might have tried this before

  9. NogginTheNog

    Bug?

    This doesn't sound like a bug to me, it's a deliberate design decision by the developers, the impact of which either they didn't understand or chose to ignore. As stated earlier though, wtf is a BROWSER doing pissing about with system wide power management features?

    1. Anonymous Coward
      Anonymous Coward

      Re: Bug?

      It looks to be both a bug and a deliberate design decision.

      The deliberate design decision was to change the system clock tick rate to 1ms when displaying flash content.

      The bug was to not change it back when it stopped displaying flash content.

      1. JeffyPoooh
        Pint

        Re: Bug?

        "The bug was to not change it back..."

        It's too bad the Windows OS doesn't have access to this setting.

        /sarcasm

  10. Lee D Silver badge

    Get rid of that awful Google Update service that constantly reinstalls itself and triggers firewalls and all sorts just to check that Chrome is up-to-date, and we'll talk.

    If you WANT to check for updates, create a scheduled task, not a service, and let me disable the damn thing.

    1. Mage Silver badge

      Get rid of that awful Google Update

      I'm still baffled why people want Google Spyware instead of an alternate browser.

      Having written programs that do mysterious things, I'd be annoyed if I couldn't do what Chrome is misusing. There is a problem though with Animation, Real Time I/O (even slowly) etc on all desktop OS. This why "serious" Industrial controllers don't run OSX, Windows or even normal Linux (a special version of Kernel is used).

      1. asdf

        Re: Get rid of that awful Google Update

        >I'm still baffled why people want Google Spyware instead of an alternate browser.

        At least on Linux its the only right out of the box easy way to run flash properly sandboxed, out of process and up to date (only Chrome has latest flash at all on Linux as its included by default which the more open sourcey projects can't do). Sadly still plenty of things on the intertubes run on flash (believe Hulu and Amazon video service do for example) and screw installing the flash plugin so every program can misuse it outside a sandbox.

        >This why "serious" Industrial controllers don't run OSX, Windows or even normal Linux (a special version of Kernel is used).

        Yes generally a RTOS. It is an IT site so don't have to assume everyone is an HR retard like at work, jk ;)

        1. asdf

          Re: Get rid of that awful Google Update

          >only Chrome has latest flash at all on Linux as its included by default which the more open sourcey projects can't do

          What I meant to say is the flash plugin for Linux is stuck at version 11 except for Chrome which comes with version 14 or whatever is latest built in and doesn't require installing a plugin any program can abuse.

  11. Anonymous Coward
    Anonymous Coward

    It was bad that no one knew about it!

    Weird that, unlike my IE that kept on running one CPU at 100%.

    You also need to bear in mind that both IE and Safari dislike HTML 5.

    Opera, Firefox and Chrome all make it close to the it works '500' mark.

    Personally IE has never been worth the security risk.

  12. psyq

    This is a Windows API problem

    Windows, even since pre-NT versions (Win95 & Co), allowed usermode applications without admin rights to change the system quantum ("tick" duration).

    By default, quantum in NT kernels is either 10ms or 15ms, depending on the configuration (nowdays it tends mostly to be 15ms). However, >any< user mode process can simply request to get this down to 1 ms using an ancient "multimedia" API.

    Needless to say, in the old days of slow PCs etc. this was used by almost all multimedia applications since it is much easier to force the system to do time accounting on 1ms scale than to do proper programming.

    For example, idiot thinks he needs 1ms time precision for his timers - voila, just force the >entire OS< to wake the damn CPU 100 times a second, do all interrupt processing just to wake his ridiculous timer because the developer in question has no grasp of efficient programming. In most cases, it is perfectly possible to implement whatever algorithms with 10/15ms quantum, but it requires decent experience in multithreaded programming. This, of course, is lacking in many cases.

    Only a very small subset of applications/algorithms need 1ms clock-tick precision. However, for those, system >should< ask for admin righs, as forcing the entire OS to do 10x times more work has terrible consequences for laptop/mobile battery life.

    Microsoft's problem is typical: they cannot change the old policy as it would break who-knows-how-much "legacy" software.

    1. psyq

      Re: This is a Windows API problem

      I meant wake up the CPU 1000 times in a second.

      1. Truth4u

        Re: This is a Windows API problem

        Interesting, what is the API and the boundaries for this tick value?

        I want to try setting it to various values between 1ms and the max to see what happens to different applications and stability.

        1. psyq

          Re: This is a Windows API problem

          It is timeBeginPeriod API:

          http://msdn.microsoft.com/en-us/library/windows/desktop/dd757624(v=vs.85).aspx

          On most systems today you can get down to 1ms. But this has a high cost associated with the interrupt overhead.

          1. Lee D Silver badge

            Re: This is a Windows API problem

            What bloody idiot allowed that to propagate through to modern Windows?

            And what kind of damage would happen if Windows just said "Okay" and then ignored the new timer functions and/or returned TIMER_NOCANDO (the most colloquially named timer constant I've ever seen)?

            MS should have deprecated that call or at least said it was "not guaranteed" 10 years ago.

            1. psyq

              Re: This is a Windows API problem

              It would not work. There are too many applications written in the crappy way with pieces of code working properly only with decreased time quantum. Their time accounting would be fcked and the results would range from audio / video dropouts to total failure (in case of automation / embedded control software).

              For example in all cases where code expects timer to be accurate to the level of, say, 1ms or 5ms. Too many multimedia-related or automation-related code would be broken.

              It is sad, but true. Microsoft should never allow this Win 3.x / 9x crap in the NT Win32 API but I suppose they were under pressure to make crappily written 3rd party multimedia stuff to work on NT Windows flavors, otherwise they might have problems with migrating customers to NT.

              Of course, nowdays (since Vista) there are much better APIs dedicated to multimedia / "pro" audio, but here the problem is the legacy.

              At least, Microsoft could have enforced API deprecation for all software linked as NT 6.x so that this terrible behavior could be avoided for new software. But that, too, is probably too much to ask due to "commercial realities". Consumer's PC would suck at battery life, nothing new here :(

              1. John Brown (no body) Silver badge
                Devil

                Re: This is a Windows API problem

                But that, too, is probably too much to ask due to "commercial realities".

                ...and yet MS can be so autocratic when they choose to be, eg lack on Win8 start menu, ribbon etc.

                Considering their market clout, they could have made API changes which pissed off developers at almost any time and the devs would have just had to suck it up. At least since Win98 MS has been the defacto desktop standard by almost any sane measurement. And before the dowvoters leap to action, I rarely use Windows myself following the one tru Daemon, FreeBSD :-)

                1. Anonymous Coward
                  Anonymous Coward

                  Re: This is a Windows API problem

                  The difference is ribbon, tiles, etc. don't break applications. API changes do. And customers may have applications they can't replace based on the old API, and sometimes they could be very important customers. If you read Raymond Chen's 'The old new thing' blog, you could find a lot of horror stories about badly implemented applications (up to reverse engineering code from libraries soon replaced they should have not used, or using unused reserved parameters later really used) - sometimes MS was forced to support those apps because used or developed by customers it cant say 'go to hell'.

                  When your software costs nothing is much easier to tell it, when your business is built on paying customers not so easy.

        2. John Brown (no body) Silver badge

          Re: This is a Windows API problem

          "Interesting, what is the API and the boundaries for this tick value?"

          +/- 100ms. Just in case you need the OS to go backwards.

  13. ElNumbre

    Poor Bug Fixing

    In my experience, this is quite common in regard to Google fixing bugs in their products. I've added my voice to reported problems in Chrome, Android, Mail amongst others, but because they're not security related or service affecting, they seemingly get ignored. I can only presume Big G don't task their developers with fixing problems, only inventing new things. And lets face it, fixing exiting broken code isn't as fun as contributing a new widget.

    Plus, Google products seem to experience entropy worse than many other - they start off really usable, fast and lightweight, and gradually erode into a big brown dysfunctional mass.

    1. keithpeter Silver badge
      Windows

      Re: Poor Bug Fixing

      "[$COMPANY | $OPEN_SOURCE_PROJECT] products seem to experience entropy worse than many other - they start off really usable, fast and lightweight, and gradually erode into a big brown dysfunctional mass."

      I thought that was a general rule, you know, like the 2nd Law? Perhaps some teams can stave it off a bit longer (*BSD, Slackware) but it gets them all in the end.

      There is a school of thought that suggests the 'big....mass' actually contains a lot of information about the problem being modelled by software.

      http://www.joelonsoftware.com/articles/fog0000000069.html

      1. Jamie Jones Silver badge
        Happy

        Re: Poor Bug Fixing

        What is the 2nd law?

        Is that the one that says all complex programs evolve to the point where they can read mail?

        1. Blue Pumpkin

          Re: Poor Bug Fixing

          The first law of thermodynamics says you can't win.

          The second law says you can't break even.

          And the third says you can't quit the game.

          1. keithpeter Silver badge
            Coat

            Re: Poor Bug Fixing

            http://blogs.msdn.com/b/oldnewthing/archive/2014/07/18/10543324.aspx

            Another example of what application developers can do.

            2nd Law is of Thermodynamics as others have pointed out.

            I'm off out while it is not raining

          2. dajames
            Boffin

            Re: Poor Bug Fixing

            The first law of thermodynamics says you can't win.

            The second law says you can't break even.

            And the third says you can't quit the game.

            I've always heard that stated as:

            1. You can't win, you can only break even.

            2. You can only break even by quitting the game.

            3. You can't quit the game.

            (where quitting the game equates to reaching absolute zero)

            1. KayKay

              Re: Poor Bug Fixing

              Hotel California

        2. dajames
          Angel

          Re: Poor Bug Fixing

          ... the one that says all complex programs evolve to the point where they can read mail?

          So, the NSA is a really complex program? I suppose that figures ... they certainly seem to give complexes to a lot of other folk.

      2. Tom 38

        Re: Poor Bug Fixing

        There is a school of thought that suggests the 'big....mass' actually contains a lot of information about the problem being modelled by software.

        http://www.joelonsoftware.com/articles/fog0000000069.html

        Consider the source... Joel makes a lot of money by spreading FUD about OSS and selling closed source alternatives.

    2. Mark 85

      Re: Poor Bug Fixing

      Let's face it, their policy with the employees is "spend time, develop something new and exciting (and profitable)". I'm sure if a developer told his boss he spent his/her time on fixing bugs they would get no points toward raise/promotion/etc.

      And since they are "investor" driven now and need to "return shareholder value" <cough>, bug fixes are a cost center that needs to be minimal and not a profit center.

  14. Anonymous Coward
    Anonymous Coward

    Not surprised

    Though I never realised it was a bug. With Mavericks I moved to using mostly Safari, on my 2011 MacBook Air, and batter life improved. On the corporate Dell from He'll, I still use Chrome but it's a hog. On Linux I avoid Chrome (find Firefox is nicer to use)!

    I'd put the issues down to the fact Chrome sandboxes each tab though, rather than a battery sucking bug!

    Erm, posted from Chrome, on the corporate Dell From He'll, complete with simulated auto correct errors.

    Anon - calling beloved employer's laptop the Corporate Dell From He'll is probably a sacking offence.

  15. Conor Turton

    Chrome sucks the life out of a Macbook Pro

    OS X may not be particularly affected by this specific bug but Chrome seriously puts a dent in battery life on a Mac to the point I use Safari on battery.

  16. Anonymous Coward
    Anonymous Coward

    Havn't been swayed

    by all the talk of how good Chrome is, still prefer Firefox for just being sensible.

    I set up a local single letter domains on a local server for various shared links, firefox will observe it as an address but good old Chrome generally wants to go look for n.com out in the real world, "I couldn't find n.com do you mean this advert, this advert or this advert?" no I mean observer the local network settings, local DNS and get the hell out of my way.

    (sorry probably said this before)

    A user at work went the chromebook route and installed lots of test apps (or whatever the addons are called) on his identity then signed in like that work, suddenly loads of issues with an otherwise sound CAD workstation. After the "IT Problems" moaning I cleaned it up and said lets leave Chrome off this time, all good and after a while we agreed to try Chrome back on but not to log in with the "all that home crap" profile, strangely similar issues very soon.

    New workstation (same OS and software) no chrome just firefox and now it's all good, old worksation new OS install to another user, no chrome all good, I see a pattern here and I can see no reason to encourage Chrome use in the workplace over Firefox.

    Chrome is becoming a virtual machine install for advertising not a browser.

    1. Salts

      Re: Havn't been swayed

      Firefox user also, though my main reason is that by using firefox, mozilla gets paid to keep competing with chrome by google, that, just appeals to me.

      That said I also find firefox a much better browser, it just works better for me.

  17. Fihart

    Google is listening ?

    If enough people dump Chrome and answer the exit questionnaire thoughtfully provided by Google, maybe things will get fixed. Note that the questionnaire also mentions privacy and app store so I mentioned worries about irrelevant permissions demanded by some apps.

    I suspect that Google do monitor uninstall reasons -- they are an information company above all.

    1. DavidRa
      Joke

      Re: Google is listening ?

      Quote: If enough people dump Chrome and answer the exit questionnaire thoughtfully provided by Google, maybe things will get fixed.

      Hi there, hope you're well. My name is Billy Bob Joe. I saw your comment and I have some excellent waterfront property going for a steal in Florida. I need to sell ASAP for personal reasons.

      But seriously, I have exactly zero confidence that those results go anywhere but the bit bucket. I strongly suspect they don't want intelligent, thoughtful users - those guys are a pain! They keep pointing out our privacy and security gaffes, and then worse - they tell ORDINARY PEOPLE! We can't have ORDINARY people understanding our data slurping!

  18. Scaffa

    The article may say it is fixed on OS X, but merely opening Chrome seems to halve the battery life on a Macbook Air.

    1. ThomH

      Don't even get me started on Google software. Of the current Android development stack, Android Studio uses "significant energy" even when left completely untouched for a prolonged period; the Android device simulator (with Intel HAXM, admittedly) does not even when in use.

      So: text editing is apparently very costly, but running a whole other virtualised OS is quite battery efficient.

    2. Anonymous Coward
      Anonymous Coward

      Do you often have Chrome open with no web pages?

      Because that's the only way you can know that Chrome is wasting your battery power vs. whatever web page(s) you have open.

  19. clriis

    I've got problems regarding the validity of this story. If this problem have existed for 10+ years, the number of users potentially involved (100 of millions of laptop users, many of them decision takers and techies) would have escalated the issue long ago. Otherwise it must be a trojan horse inflicted by Google.

    1. Anonymous Coward
      Anonymous Coward

      You're ignoring the severity of the issue. If it only decreases battery life by a few percent then how would anybody know to complain about it?

  20. Patrick R

    Windows, not on Linux and OS X machines

    What about Chrome OS then ?

  21. Anonymous Coward
    FAIL

    Use of the timeBeginPeriod / timeEndPeriod Windows API calls is bad coding practice nowadays. These calls have existed since Windows 3.1. Google should know better.

  22. JeffyPoooh
    Pint

    CS = 'C' Students

    Chrome under Android on a Nexus tablet crashes all the time. 100% Google.

    Even their Gmail spam filter is failing lately, eating non-spam messages.

  23. Crisp

    Now we know how the Google bug ticketing system works.

    1) Find bug

    2) Post bug in an article to Forbes

    3) Google fixes the bug.

  24. Fihart

    More Chrome issues.

    To quote a friend:

    "wouldn't advise anyone to use Chrome unless they had 8mb of Ram! tho a workaround is the Chrome extension "OneTab" which puts all your open Chrome tabs into just one tab, greatly minimising RAM usage - even so Chrome hooks really hurt available RAM!"

    1. Anonymous Coward
      Anonymous Coward

      Re: More Chrome issues.

      Or install the great suspender app if you have a habit of running multiple tabs open for days/week rather than book marking them.

    2. Anonymous Coward
      Anonymous Coward

      Re: More Chrome issues.

      I use Chrome on my MacBook with 2GB RAM and it runs fine, minimal memory pressure. I also use it on my HTPC with 2GB RAM which runs Windows, usually not many tabs open (HTPC) but also typically only ~1GB RAM is being used at any given time. My main machine is a Mac with 4GB RAM and I have all sorts of tabs and other software open and it's rarely under any memory pressure.

  25. Gis Bun

    More sloppy programming from Google. Makes Apple look like the second best programmers in the world. :-)

  26. Bucky 2

    Child's Browser

    I once made the comment that Chrome was a "Child's Browser." I don't know where it came from. It was an off-the-cuff, meaningless remark.

    But it made a junior programmer here SO MAD I started using the term over and over again. You know, for sport.

    I wish he was still around... I'd forward this article to him.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like