Microsoft: Internet Explorer 9's "industry leading" energy efficiency

Here's something to think about: when you ask some of your fellow geeky friends what they find best about the recently released Internet Explorer 9, see what responses you get back. You will most likely hear the words "fast," "light," or even "elegant" thrown around. Some of your acquaintances may dispute those claims. But one thing that Microsoft is not emphasizing much in their IE9 marketing blitz, but are clearly leading their competitors on, is energy efficiency. If you're a energy-conscientious laptop user, take note! If you think your energy bills could shave a few pennies this month, also take note.

On the IE blog, three Internet Explorer team members shared the results of several energy tests they conducted on the five most popular browsers available for Windows. They tested the amount of power the browsers consumed when idling on about:blank pages, when loading news sites, and working through two example HTML5 applications offered on the IE Test Drive site - "Galactic" and the ever-so-recognizable "FishIE Tank."

The results are interesting. For the about:blank test, most of the browsers did not use much more power compared to a system on idle. The one anomaly in that test was Opera, which consumed the most power on idle - the reason we'll get to in a bit. For the news site tests, Firefox, Safari, and Internet Explorer fared well, while Chrome and Opera caused  more noticeable power spikes. Firefox and Internet Explorer fared pretty well in the HTML5 applications thanks to GPU acceleration taking some load off the CPU.

The question remains: besides the lack of GPU acceleration in some of these browsers, what are they doing wrong that is causing increased power usage? In the case of Opera, even while the browser was relatively idle, it lowered the system timer resolution from the default of 15.6 ms (for Windows 7) to 2 ms, which improved responsiveness but prevented the CPU from entering lower power states. Internet Explorer, on the other hand, keeps the system timer resolution at 15.6 ms while the user is running on a battery, and only lowers the resolution when the system is plugged back in. Users can check the current system timer resolution by running Sysinternals' ClockRes utility.

In the meantime, this may be a cue for browser vendors to focus their attention on improving power efficiency, in addition to the current speed race amongst browsers.

Image Credit: IE Blog

Report a problem with article
Previous Story

Windows Phone 7 predicted to significantly outsell iPhone by 2015

Next Story

TechSpot: Latest in Notebooks - First Half of 2011

33 Comments

Commenting is disabled on this article.

Wait, so chrome kills my battery faster but consumes the same number of watts as firefox4 or ie9?

Ned said,
Wait, so chrome kills my battery faster but consumes the same number of watts as firefox4 or ie9?

I didn't see that about:blank. Makes more sense now.

Interesting test...I wouldn't have thought it was such a large difference...
I'll have to try it with Linux sometime...Well...With the browsers that work on Linux anyways.

That's because IE's components are apart of windows and are already loaded. Did you try the same test on OSX or under Wine in GNU/Linux? Thought not...

Flawed said,
That's because IE's components are apart of windows and are already loaded. Did you try the same test on OSX or under Wine in GNU/Linux? Thought not...

How would that matter? They cant test IE9 on OS X, but they can test Opera, Safari, Chrome and FF4. But that wont matter. Microsoft was testing IE9, they cant do that on OS X. If you want to try the test then do it yourself.

Personally I use Chrome when on I'm charging and Safari when I'm on battery under OS X. I notice a rather large (30+min) battery increase when using Safari over Chrome on battery.

Flawed said,
That's because IE's components are apart of windows and are already loaded. Did you try the same test on OSX or under Wine in GNU/Linux? Thought not...

Preloading components has little effect on CPU usage... we're not talking about loading times here.

Now that's a statement I agree with, loading same pages with IE9, firefox 4, opera or chrome will always result in IE9 abusing less CPU power, chrome is the worst of them in power usage.

dangel said,
I see no change in clock resolution with Opera running or without it. Daft.

You might have another app in the background that changed it before Opera did.

Lol, I'm not switching browsers to save a few pennies. Even if I was a Opera user, a few pennies really isn't a... don't know the word.

david said,
Lol, I'm not switching browsers to save a few pennies. Even if I was a Opera user, a few pennies really isn't a... don't know the word.

It's not just about saving money. It's also about being out and not having to charge as quickly which could be important in situations when you simply can't charge because you don't have access.

david said,
Lol, I'm not switching browsers to save a few pennies. Even if I was a Opera user, a few pennies really isn't a... don't know the word.
He is probably laptop-less, living in a desktop world.

I didn't know about the increased power usage with Opera, but I figured hardware acceleration would improve battery performance so no surprise there.

Anaron said,
I didn't know about the increased power usage with Opera, but I figured hardware acceleration would improve battery performance so no surprise there.

There is also how much the browser is 'quiet' which then allows the CPU to fall into lower power levels; if the browser is constantly polling and making noise then the CPU never goes into the lower power state and thus more power is used. I'm not surprised that Safari didn't perform as well considering that running on Windows is more of an after thought. As for Chrome - tab process seperation is always going to be heavier on the CPU but apparently Webkit2 goes about it in another way which, from my experience so far with Lion, hasn't impacted on the battery life. IMHO I'm happy to sacrifice some battery life for greater stability.