Does Windows Really Have Such Terrible Battery Life?

Let's not conflate constructive criticism with hyperbole.

  • Share
  • Read Later
Microsoft, Apple

Jeff Atwood of Coding Horror is asking an important question, one that’s been glossed over in tech-dom for years: why does Windows have such poor battery life on mobile devices compared to the competition?

No one seems to know. Atwood asked one of tech’s luminaries on the enthusiast side, Anand Shimpi of Anandtech, but says Shimpi was as puzzled as him. You’d expect Microsoft software engineers to know, but then you’d also imagine they’d have diminished the discrepancies between Windows-based devices and others by now.

Instead, Windows battery life seems to be headed in the opposite direction. Atwood cites Anandtech’s review of the Surface Pro 2, which includes a battery benchmark chart that shows the Surface Pro 2 clocking just below seven hours of usage browsing the web on Wi-Fi. What’s so bad about seven hours? Not much, and Atwood agrees that seven hours is pretty nice. But how does the Surface Pro 2’s 42 Wh (Watt hours) battery stack up against a device with a similar spec battery?

Tapping another Anandtech chart — this one of the site’s review of Apple’s 2013 13-inch MacBook Air — Atwood notes the Air, with its slightly lower 38 Wh battery rating, manages over 11 hours of battery life while web browsing on Wi-Fi.

Thus, as Atwood puts it (his emphasis):

That means the Air is somehow producing nearly two times the battery efficiency of the best hardware and software combination Microsoft can muster, for what I consider to be the most common usage pattern on a computer today.

I don’t own a Surface and don’t intend to buy a Surface, but I do have a Retina MacBook Pro, and I’ve run various versions of Windows in Boot Camp mode since Apple made the leap from Power PC to Intel, mostly to play games like StarCraft II and Guild Wars 2 without OS X’s performance penalties. In all cases, no matter the version of Windows or OS X, Windows battery life while doing the sort of work I’m doing right now — emailing, browsing the web, light image editing — is significantly less than OS X’s, and I’ve tallied the difference in hours, not minutes. Even Windows 8 does the operating system’s reputation as a battery-vampire (compared to Apple’s OS X) no favors.

Is it because Apple has greater control over its hardware, while Microsoft has to support legions of potential hardware combinations, frustrating its ability to optimize energy distribution? Maybe, but Extreme Tech says not so fast, noting the Surface Pro 2 was designed in-house, just as Apple products are. What’s more likely, argues ET, is that Microsoft’s engineers either aren’t as good at or as focused on improving battery life. Take your pick, but for me, that’s just another way of saying “Who knows?”

What’s so bad about seven hours? With Intel’s Haswell chips providing devices a huge helping hand, power-management-wise, does it really matter that you’re getting seven hours next to someone else’s 13? Extreme power users notwithstanding, do you sit for more than seven hours in a coffee shop browsing the web? And even if that seven hours trades down to four or five doing work in light apps, do you really need more than that between charges?

That’s the question you have to answer (and no review can account for): Is seven hours for web browsing enough? And if it is, does anyone care that another company offers more? Should we be calling Windows battery life “terrible” if — and I repeat if — it passes muster for most use case scenarios? In short, I love that Atwood’s asking the question, and I hope he keeps asking it, I’d just like to see him tweak the descriptor “terrible.” Seven hours is hardly terrible, whether taken by itself, or compared to something that offers twice as much.

Yes, there’s an environmental angle here I’m not forgetting: Even if the new Surface tablets exceed your battery usage requirements, you’re going to use more juice, on balance, than you would with an Apple device. You’ll also cycle charges more frequently, thus your overall battery life is going to be theoretically less, assuming you’re going cordless with one device as often as the other. Regardless of your particular needs, then, it would behoove Microsoft to retool future versions of its flagship operating system to rectify this shortcoming and at least bring Windows up to par with Apple’s operating systems.


If you are going to compare eggs with eggs, then I would suggest that Windows 8 or Mac OS X be installed on the same hardware and then power consumption measured and compared?


Stupidity rises again. Do some friggin' research.

1. The Surface Pro 2 uses the more powerful Y-series i5 while the Air uses the less powerful, lower voltage, lower clocked U-series i5.

2. The Surface Pro 2 also has a significantly higher resolution screen. This is not hard to figure out.

There's your answer right there.


In this case, my guess is that there are two culprits that contribute to the battery life difference: legacy software on Windows and display. 

Most power is spent on CPU, GPU, display, and wireless. 

In this case, the Surface Pro has a higher resolution and a smaller screen. To support the same brightness, more backlight (hence power) is needed for higher pixel density than for lower (Air11" - btw, you have a typo - it's the 11" MBA that has 38 Wh battery, not 13").

Then there is CPU, GPU and wireless. In order to save power, access to CPU and to wireless must be coalesced to allow for downtime when power is conserved. This is done in order of milliseconds, so it's not noticeable. The legacy software on Windows doesn't care about coalescing, and may even be using legacy API that needlessly keep the wireless card from shutting down. And there may be reasons why at the driver level this is not possible without breaking legacy code.

So... this is my somewhat educated, but not at all expert, guess.


I would guess that Microsoft's monopoly, that enabled them to force hardware manufacturers to write drivers for them is finally backfiring on them. Now instead of benefiting, they are having to pay the price of loss of control of their own drivers. Android (Linux) have suffered for years, often having to reverse engineer drivers, but now they benefit because most drivers are part of an open source kernel. Of course Apple has always had full control since they manufacture the whole product, the hard and soft parts...