Chrome Beats Edge in New Battery Life Tests

Microsoft really, really wants you to use its Windows 10 browser, Edge. The company has come under fire for its relentless attempts to push consumers towards its own software, even going so far as to toss pop-up adds into the system tray when you're running different browsers.

The company's claims have largely gone unchallenged, in part because browsers like Chrome have long been known to be resource hogs. Chrome opens every new tab as a new process, which tends to increase memory usage. Linus Tech Tips recently took four identical machines from Dell and put them through a battery of independent browser tests to see how Microsoft Edge would hold up against Opera, Chrome, and Firefox. The site tested the systems before and after the Creators Update, and took pains to attempt to duplicate Microsoft's own test environment.

The results are interesting for several reasons. First, they show tremendous variation between laptops. Machine 3 somehow manages to turn in the worst scores in every browser but Chrome, where it nearly leads the pack. The LTT team notes that they controlled for variables like system updates, telemetry transmissions, and "value added" software before beginning these tests, and that they used the same set of software on every system. Even so, averaged across all the systems, Edge has worse battery life than Chrome or Opera. LTT also attempted to duplicate Microsoft's Vimeo test, but again, battery life didn't come out with Edge on top.

I have no issues to raise with Linus Tech Tips' testing or test design, at least not as far as what's in the video (below). It's possible some minor variation in system configuration impacted these results, but it's also possible that we see different results from different systems because of variation in hardware power consumption. One thing we often talk about when we review system components is that different hardware can exhibit different results when it comes to power consumption or overclocking. We could be seeing some of those variables at work here.

The only other thing I'd add is that measuring power consumption and battery life can depend on subtle differences between applications that aren't always readily apparent. I've been testing an Alienware laptop with an OLED panel (more on that coming this week). One thing I noticed when I took the laptop on a trip was that battery life was far better when using Microsoft's built-in Windows 10 video player than when using either VLC or MPC-HC. I tweaked settings in both applications, I switched between various hardware decode methods, and I checked which GPU (Intel's onboard graphics or the Nvidia GTX 1060) was handling the decode. Putting the laptop in Power Saving mode helped, as one would expect, but also made video more prone to stuttering, no matter which hardware decode method I chose. Microsoft's included Windows 10 video playback, on the other hand, had no problems at all with smooth playback in that mode, and it offered the best battery life overall.

My own experience is only tangentially related to Linus Tech Tips' findings, but it does serve to highlight an important point. We like to think of battery life as a static number, perhaps with one value for "Idle" and another value for "Load." But real life isn't that simple. How systems are configured, even at the most basic level, can impact how battery tests resolve. We'll be curious to see if and how Microsoft responds to these findings from LTT. But it wouldn't surprise me if there were subtle configuration differences (fair or otherwise) that account for the difference in results. In my personal experience, these types of tiny variables can often account for major differences.