Browser Face-Off: Battery Life Explored
by Jarred Walton on September 11, 2009 6:00 PM EST- Posted in
- Laptops
Netbook Browser Battery Life
The first two laptops are representative of a large portion of the mobile market. Both processors are reasonably fast, you get 4GB RAM, and battery life is acceptable if not great. Discrete graphics would improve performance in gaming applications by a substantial amount, but it would also eat into the battery life, which is why most laptops utilize integrated graphics. Our last laptop for the browser battery life testing represents the last major chunk of the mobile market where battery life would matter: netbooks. The ASUS 1005HA represents the current state-of-the-art in netbooks, with an Intel Atom N280 processor, 1GB memory, a 160GB hard drive, and a 10.1" LCD. It uses the Intel 945GSE + ICH7MU chipset with integrated GMA 950 graphics.
Unlike our testing for the 1005HA review, we set the Super Hybrid Engine to "High" instead of "Auto", so the N280 won't underclock on battery. This reduces battery life by around 6%. Since the 1005HA runs XP instead of Vista, we also chose the "Portable/Laptop" power profile rather than the "Max Battery"; the reason is that the netbooks tend to run too slow when set to "Max Battery" and the gains aren't as substantial as with faster laptops. We measured a 7.5% increase in battery life by switching to the "Max Battery" profile, if you're interested.
| ASUS Eee PC 1005HA Specifications | |
| Processor | Intel Atom N270 or N280 (Tested) N280: 1.66GHz, 512KB L2, 45nm, 667FSB |
| Chipset | Intel 945GSE + ICH7MU |
| Memory | 1x1024MB DDR2-533 @ 4-4-4-12 Timings |
| Graphics | Integrated Intel GMA 950 |
| Display | 10.1" Glossy LED-Backlit ~16:9 WSVGA (1024x600) |
| Hard Drive | 2.5" 160GB 5400RPM 7MB (Hitachi HTS543216L9SA00) |
| Networking | Atheros AR8132 Fast Ethernet Atheros AR9285 802.11n WiFi |
| Audio | Realtek AL269 2-Channel HD Audio (2.0 Speakers with headphone/microphone jacks) |
| Battery | 6-Cell 11.25V, 5600 mAhr, 63.0 Whr |
| Front Side | None |
| Left Side | Heat Exhaust Kensington Lock 1 x USB 2.0 VGA AC Power Connection |
| Right Side | SD/MMC reader Microphone/Headphone Jacks 2 x USB 2.0 100Mb Fast Ethernet |
| Back Side | None |
| Operating System | Windows XP Home SP3 |
| Dimensions | 10.31" x 7.01" x 1.02"-1.44" (WxDxH) |
| Weight | 2.8 lbs (with 6-cell battery) |
| Extras | 1.3MP Webcam Super Hybrid Engine (software over/under clocking) Available in White, Black, Blue, and Pink |
| Warranty | 1-year standard ASUS warranty (USA) Extended warranties available |
| Price | Black 1005HA-PU1X-BK starting at $381 |
Unlike the other laptops, all of the browsers offer almost the same battery life on the 1005HA. Google Chrome comes in first this time, offering 2.5% more battery life than IE8. There's only a four minute gap between IE8 and Opera 10, a difference of less than 1% total. AdBlock doesn't help Firefox this time, most likely because of the amount of processing it takes to parse HTML and perform regular expression matching.
The gap between first and sixth with a netbook looks to be 3.5%, so if you're using a netbook you can select a browser based on other features, though Chrome does come in first by almost 15 minutes. The major exception is that Apple's Safari 4 web browser again comes in last, again by a large margin. Chrome managed 36% more battery life on the 1005HA than Safari, and Flash content within Safari 4 used a significant amount of CPU time. Safari seems to want more CPU power than the Atom can provide, with the result being the processor is often near 100% utilization for significant amounts of time on Flash-heavy sites.
Facebook
Twitter
RSS
76 Comments
View All Comments
Granseth - Saturday, September 12, 2009 - link
Opera 10 has something called turbo mode to accelerate netsurfing on a slow connection. Could be interesting to know if that would help or hinder batterylife toosolipsism - Saturday, September 12, 2009 - link
Could you do comparative tests of browsers under OS X? If I can save 10% power while forced to use my battery I’ll likely do it based on your results. Also, do using plugins like ClickToFlash in Safari on OS X have any overhead that negate any power savings from turning off Flash?orthorim - Thursday, September 17, 2009 - link
YES, please test Mac too. I run OS X so the tests on Windows are meaningless to me. Presumably, Safari on OS X will be much more optimized than on Windows, and Firefox will be totally different as well.I'd like to see Safari + ClickToFlash because that's what I am running every day. And Firefox with and without Flashblock
Flash should make a huge difference on OS X because it's so poorly implemented by Adobe. Even simple Flash animations use lots of CPU on OS X, whereas on Windows they are hardly noticeable. I want to see how much ClickToFlash/Flashblock help, and which browser is the best.
Firefox
Chrome Beta
Safari
Camino
Opera (?!)
Thanks!
ltcommanderdata - Saturday, September 12, 2009 - link
If you do run browser battery life tests in OS X, can you make sure to try Safari in both 32-bit and 64-bit modes? I don't expect the browser itself to affect battery life that much between the 2 modes, but 64-bit Safari is capable of playing Flash content. I believe 64-bit Safari is still using the standard 32-bit Flash plugin, but as a separate process, giving crash stability, and using InterProcess Communication to connect the 2. I'm thinking the IPC link in 64-bit mode may use more CPU cycles reducing battery life.JarredWalton - Saturday, September 12, 2009 - link
Talk to Ryan and Anand - they're the Mac OS X people. I don't have a Mac at my place, or I'd be happy to run the tests. :-)GeorgeH - Saturday, September 12, 2009 - link
If Anand or Ryan decides to run browser tests on a Mac, they might also want to redo the test where Anand found that OSX 10.5.7 had vastly superior battery life to Vista. Given how poorly Safari under Windows did here, the methodology of that test becomes suspect.In that article Anand found that Vista running Safari offered ~75% of the battery life of OSX running Safari. Here you find that Safari under Windows offers ~75-80% of the battery life of IE (and most other "typical" Windows browsers.) Given the similarity of those numbers, perhaps Anand should have concluded that Safari/Vista is inferior to Safari/OSX, and not that OSX was the superior OS in general.
Article:
http://www.anandtech.com/mac/showdoc.aspx?i=3582">http://www.anandtech.com/mac/showdoc.aspx?i=3582
solipsism - Saturday, September 12, 2009 - link
I think that when doing OS comparisons that IE8 on Windows v. Safari 4 on OS X, then the same version of Firefox on each OS would be the most apt.foolsgambit11 - Sunday, September 13, 2009 - link
Maybe not. It's difficult to tell how optimized FF would be for each platform, while testing each OS maker's proprietary browser would ensure they've both done their best to ensure maximum optimization. Of course, ideally, the benchmarks would also test FF to see how well it does on each platform.JarredWalton - Saturday, September 12, 2009 - link
Anand's testing didn't use heavy Flash content, and that appears to be the primary issue with Safari 4 under Windows. If I switch to static images and text without Flash, as you would expect CPU usage drops to nothing after the pages finish loading.GeorgeH - Saturday, September 12, 2009 - link
That's a very valid point, but doesn't quite address what I was trying to get at.What I was trying to say is that by simply using "poorly optimized" (or however you want to phrase it) software on Vista, your battery life can drop significantly. Anand tested OS battery life by running two programs that were designed for OSX, not Vista, and found that the Vista performance deficit was almost exactly the same as the one you found by running "poorly optimized" software.
That the cuplrit in your case was almost certainly Flash (which Anand didn't run) is relevant, but doesn't address the idea that by running software in an environment that it wasn't designed/optimized for it is possible to significantly impact your battery life. To quote myself from the other article:
"I doubt it will make much difference, but in the interest of fair play I too would like to see the tests [OSX v. Vista] redone [using standard Vista programs]."