18

I was reading an article on Forbes that says Google Chrome might be draining my laptop's battery because it brings the system into a mode that uses a tick rate of 1ms instead of 16ms, and is using up to 25% more battery power in the process. (And apparently, it doesn't leave that mode like it's supposed to until you close the application, instead of when you're done with the high performance units of work.)

I understand that a shorter time slice means more context switching, which would cause an increase in CPU usage, in turn causing increased battery consumption. But it doesn't seem to be sourced anywhere that it uses a whopping 25% more from that one setting.

Is there evidence to support that the decrease in time slice duration causes the dramatic battery use increase claimed in the article?

2
  • According to ZDNet, Microsoft also claims "up to 25%". The Intel link has some relevant charts, but it doesn't give an exact number for "this is how much more power it takes" that I can see. FWIW, Google says this is a feature, not a bug, but will fix it anyway.
    – Is Begot
    Aug 1, 2014 at 15:07
  • This blog post might be of interest: Windows Timer Resolution: Megawatts Wasted
    – and31415
    Aug 2, 2014 at 17:49

1 Answer 1

17

Is there evidence to support that the decrease in time slice duration causes the dramatic battery use increase claimed in the article?

Yes.

Quoting a technical paper released by Microsoft as part of their hardware developer documentation, Timers, Timer Resolution, and Development of Efficient Code:

If the system timer interval is decreased to less than the default, including when an application calls timeBeginPeriod with a resolution of 1 ms, the low-power idle states are ineffective at reducing system power consumption and system battery life suffers.

System battery life can be reduced as much as 25 percent, depending on the hardware platform. This is because transitions to and from low-power states incur an energy cost. Therefore, entering and exiting low-power states without spending a minimum amount of time in the low-power states can be more costly than if the system simply remained in the high-power state.

According to PC world, a global reputable computer magazine published monthly by IDG, Google has confirmed the issue:

In any case, the new reports have finally gotten Google's attention. In a statement to PCWorld, the company noted that the bug has been assigned internally, and that the Chrome team is working to fix it—though only after Morris shined a spotlight on the issue. The long-standing bug report has been bumped up to priority one.

More details:

The bug mentioned in Forbe's article, has been reported in Chrome's Chromium open-source project since September 29, 2012. This bug has grabbed 7445 votes from programmers.

Apparently, Google's browser increases the system clock tick rate to 1ms as soon as you launch the browser and keeps it that way until it is completely shut-down. This means that the system's CPU needs to switch to 'wake up mode' 1000 times in 1 second, to attend to Chrome-related tasks.

Microsoft’s default setting for Windows is 15.6ms, which will wake the computer 64 times in 1 second. Because Chrome forces the system clock resolution to 1.0ms, this causes substantial power drain, since this is much more frequent than usually observed with Microsoft's default setting.

In short, Chrome is waking the processor up 1,000 times per second instead of just 64 times per second.

8
  • 2
    I do not find that quote at that link. Further, the Chrome team has only agreed to address the time-slice issue, which is completely isolated from the issue of whether or not the power consumption is increased. It merely shows that people are reacting negatively to the bug report - my question is specifically about whether or not they are overreacting.
    – corsiKa
    Aug 1, 2014 at 15:51
  • 5
    But it uses weasel words. "Up to" 25% "depending" on platform. It could be it's more like 2% on most platforms, but one or two obscure ones get 25%. Or maybe most are in the 20% range and a few peak to 25%. It's not a very convincing summary.
    – corsiKa
    Aug 1, 2014 at 17:59
  • I acknowledge that there is going to be some additional drain due to the decreased time slice duration. The question is whether it's marginal or significant. An independent source (not just the one quoted in this article, but perhaps other research into the problem area, microsoft can't be the only people interested in battery life) would be convincing, as would details about this mysterious 'fact' microsoft is publishing without seriously backing up with data.
    – corsiKa
    Aug 1, 2014 at 18:04
  • @corsiKa - If you're looking specifically for proof of the 25% figure, then you asked a bad question. You should edit it to include the quote from Microsoft's documentation, and change the question you're asking from "Does Chrome do this?" to "Under what circumstances are the numbers this bad?" Or just ask a new question.
    – Bobson
    Aug 1, 2014 at 20:27
  • I'm asking for that proof of the 25% because it was presented here as a 'fact' in an 'academic' paper that reads more like marketing than academia. My question about general power consumption and whether or not it's a significant concern is still valid and is the right question.
    – corsiKa
    Aug 1, 2014 at 20:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .