Surface Pro 2 and Windows Battery Life with Haswell

By Wesley Fenlon

Does Windows have terrible battery life? Compared to OS X, yes. But it's hard to say why.

Programmer Jeff Atwood, who runs the blog Coding Horror, has a loaded question to ask about Windows and Microsoft's new Surface Pro 2 tablet. It's also an important question: "Why does Windows have terrible battery life?" Before getting around to the why, it's worth confirming that Windows actually has terrible battery life. Does it? After all, Intel has been pushing for 10+ hour battery life from every system wearing the Ultrabook name, and Haswell chips just dramatically cut power draw compared to last year's Ivy Bridge processors. Things should be looking pretty good for Windows right now.

In the grand scheme of mobile technology, Windows battery life is terrible. Atwood's claim draws heavily on extensive testing performed by Anandtech, and that testing shows the Surface Pro 2 lagging far behind virtually every other tablet in battery life. With 6.8 hours of battery life, the only tablet it beat was the first-gen Surface Pro, which got only 4.7 hours.

In a way, comparing the Surface Pro 2 to the iPad or a Samsung tablet is unfair. It's a fully-powered x86 machine in tablet form, running much more powerful hardware than a more power efficient ARM chip. But the discrepancy remains. The iPad 4 gets nearly 10 hours of battery life. The Nexus 7 passes 12. The Nvidia Shield hits 13.5.

Even if you accept that the Surface Pro 2 is more laptop than tablet, its battery life isn't much to write home about. Then Atwood compares the Surface Pro 2 to Apple's 2013 MacBook Airs, with more data from the 11-inch and 13-inch models courtesy of Anandtech. They achieved over 11 and 14 hours of runtime, respectively.

"Let's see how the 2013 MacBook Air does, which spec-wise is about as close as we can get to the Surface Pro 2," Atwood writes. "The screen is somewhat lower resolution and not touch capable, of course, but under the hood, the i5-4200u CPU and LPDDR3 RAM are nearly the same. It's a real computer, too, using the latest Intel technology.

Image via Anandtech.

"The Surface Pro 2 has a 42 Wh battery, which puts it closer to the 11 inch Air in capacity. Still, over 11 hours of battery life browsing the web on WiFi? 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. That's shocking. Scandalous, even."

Anandtech wasn't alone in finding the Surface Pro 2's battery life relatively lackluster. Under heavy usage, Gizmodo found that the tablet performed little better than the first Surface Pro:

"We were able to squeeze three and a half hours of generating internet for Gizmodo dot com out of a single charge, juggling some 10-15 open Chrome tabs and pushing video out to an additional monitor. This is almost the same as we got out of the original Surface Pro, which is a little odd considering Intel's Haswell is such a battery saving monster. Likewise, the Pro 2 performed almost identically to the Pro in a 10-hour YouTube video test, giving up at just over two and a half hours, despite fairing better in situations that seem like they should be more intense."

But this makes sense. Haswell's strength lies in its idle power usage, and constantly running video, a dozen Internet tabs, and outputting to a secondary display is not going to show off battery longevity. This doesn't seem like a problem with Windows battery life. Other tests tell a different story.

Engadget's test puts things in a bit more perspective. Engadget writes "the battery life is actually much improved this time around. Whereas the original could barely crack four hours, this year's model lasted through six hours and 27 minutes of uninterrupted video playback (that's with WiFi on and the screen brightness fixed at 65 percent)...Still, while six hours and change might be good for a Core i5 tablet, it's a pittance compared to what you'd get from most Ultrabooks. You can blame Intel, if you like, for setting the bar so high with its Haswell chips, but at this point, we expect that even the skinniest Ultrabook will last at least seven and a half hours on a charge. Most, though, last somewhere between eight and half and 10, with one freak of nature managing 12.5. In Microsoft's defense, the Surface Pro 2 weighs nearly a pound less than those Ultrabooks we're referring to (all of which have 13-inch screens). A smaller device means a smaller battery, which means less battery life; makes sense to us."

Their battery tests with other devices backed up their claims; Engadget's battery chart showed the MacBook Air, Sony Vaio Duo 13, Samsung ATIV Book 9 Plus, Vaio Pro 13, Acer Aspire S7-392, Iconia W700, and Vaio Pro 11, all tablets or Ultrabooks, delivering better battery life.

Apple [devices are] known for great battery life, and the common wisdom is that Apple's control of hardware and software is responsible. But now Microsoft is in the same position. Shouldn't the Surface Pro 2 be beating third-party Windows machines? "Microsoft can no longer hand wave this vast difference away based on vague references to 'poorly optimized third party drivers,'" Atwood writes.

Atwood blames Windows' idle power management, and some commenters chimed in with supporting evidence, citing tons of background threads and activity on fresh installs of Windows 8 running on the surface. Others blamed applications, instead. One user wrote "Traditionally, the big killer of battery life on Windows hasn't been kernelspace or system code, it's been userland apps. Apps which engage in polling, for example, forcing the system to spin up a CPU or--worse--access disk. Load up procmon on an idle system, watch the stuff that scrolls by while you're not directing it to do anything. That's the stuff that's usually eating your battery.

Even the userland apps that Microsoft puts out vary in quality. Their helper application for Intellimouse polls, for example."

Windows clearly has an issue with using too much power, even when it isn't running heavy processes like HD video streaming.

Either way, Windows clearly has an issue with using too much power, even when it isn't running heavy processes like HD video streaming. And Atwood is right--it's perplexing that Windows doesn't do better. After all, remember timer coalescing, which Apple talked up earlier this summer for OS X Mavericks? The feature coordinates CPU and disk usage across applications and background processes, so the CPU and disk can spend more time idle, and thus draw less power. This is a new feature for OS X Mavericks--but Microsoft implemented it in Windows 7 in 2009. Four years ago.

When Windows is installed on MacBook hardware, it regularly delivers lower battery life. Many people blame this on crappy Windows drivers from Apple, but one of Atwood's commenters makes a compelling case that that's not true, writing "If you look at what the Windows boot camp driver partition actually contains it's just the totally standard (and usually several versions behind current) Nvidia/AMD graphics drivers, Intel chipset drivers that were standard at the time for that CPU, normal atheros wireless drivers, bluetooth drivers, etc...All that stuff is exactly the same as you'd get on a comparable hp probook or lenovo, etc."

He even addresses the theory that there are some power management modes only OS X can access on Mac hardware:

"A compelling argument if someone could back it up would be that there's some complicated power management modes in the PMU/battery controller that only OSX gets to touch, much like how some of the switchable graphics macbook pros can only use the high-power card on windows. If that was the case though, then i still don't see how we should be seeing such a stark difference on the integrated graphics only systems though. This gap existed even in the 1st gen plastic macbook days with the gma950."

The simplest answer, it seems, is that Windows needs to overhaul its idle battery processes, either to better reign in applications or processes running in the Windows kernel. No matter what Microsoft does, battery performance is always going to vary from one OEM to the next, as different combinations of hardware and drivers will affect Windows performance. Sony's Vaio Pro 13, for example, actually beat the 2013 MacBook Air in Watt hour performance in a light usage test, but was significantly less efficient than the MacBook Air under medium and heavy usage. The same MacBook Air running Windows performed worse in the light and medium tests, but nearly matched its OS X counterpart in the heavy test.

The Vaio Pro 13 is a good example of Windows hardware that can come close to Apple's, but there's still a gap. And the Microsoft Surface Pro 2--which Microsoft has complete control over, on both the hardware and software fronts--is the system that should be delivering the best battery performance Windows can muster. Microsoft has some work to do for Windows 8.2.