Printed from http://www.electronista.com

Study: Nexus S, Android beat iPhone 4 on web 84% of the time

updated 09:35 am EDT, Thu March 17, 2011

Blaze says Android faster on web than iPhone

(Update: doubts surface) A new study by Blaze has suggested that Android in ideal conditions can easily outperform an iPhone 4 on the web. A Google Nexus S running Android 2.3 is about 52 percent faster on average than an iPhone 4, even with the iOS 4.3 speed boost in effect. In the real world, the Nexus S would outrun the iPhone 4 on Fortune 1000 company websites 84 percent of the time, the Ottawa-based firm said.

The edge came primarily for regular, desktop-optimized sites. Both phones were in a dead heat for mobile sites, but Android pulled well ahead in desktop pages, where the Chrome-derived engine is more intelligent: it tries to download as many images from the same server as possible and can get script files for CSS, JavaScript and other code in parallel, rather than serial in iOS. Apple now has a very fast JavaScript engine in Nitro that outperforms Android 2.3's engine, but this only comes into play when on a script-heavy site or a benchmark, Blaze explained.

Virtually all of the 45,000 tests between the two devices were conducted on a known good Wi-Fi connection to eliminate the connection as a bottleneck, but the iPhone using 4.2 was tested on Bell's 3G network at low-traffic moments and showed about half a second's difference from using its 7.2Mbps HSPA access. Using iOS 4.3 might improve the load times, while using any US carrier might have decreased the peak download speeds; at nearly 6Mbps downstream, Bell's 3G speed is largely unachievable in the US on AT&T, T-Mobile, or Verizon, the carriers that offer the phones in the country.

The researchers acknowledged that conditions would vary in the field and that phones using Android 2.2 wouldn't fare as well. Since Galaxy S phones like the Vibrant are the same as the Nexus S apart from using the older OS, Blaze could test pure software limitations. Android 2.2 was about 10 percent slower than 2.3, Blaze said.

Questions remain about whether or not the performance edge is feasible for most Android users. Android 2.3 represents just 1.7 percent of devices using the OS, while 61.3 percent are still using Android 2.2. In many cases, the phone manufacturers either don't have a 2.3 upgrade ready due to OS fragmentation or have deliberately abandoned updates to reduce their support costs and steer customers to newer phones. A significant number of Galaxy S-based phones in the US still use Android 2.1.

The iPhone 5 is also expected to get the same dual-core A5 chip as the iPad 2 and may get further improvements from iOS 5, which is likely to show in the next few months. Tests weren't conducted with dual-core phones like the just-launched Motorola Atrix, but these are either too new or, in AT&T's case, have been intentionally slowed down while testing an upgrade for upload speeds.

Update: A rebuttal has claimed the testing method is flawed. The Loop notes that the same problem with slowdowns in web apps launched from the home screen would also affect the software wrapper Blaze used for the test. With the UIWebView control in use instead of Safari, the Nitro JavaScript engine, multi-threaded processing, the more advanced caching, and multithreaded code are turned off. A raw, browser-to-browser test might come in Safari's favor.

Blaze has admitted that it hadn't checked if anything had been disabled. JavaScript was only about 15 percent of the rendering work, it said, but it also didn't have an explanation for other disabled features.







By Electronista Staff
toggle

Comments

  1. pairof9s

    Mac Enthusiast

    Joined: Jan 2008

    +4

    Awesome!

    I'm selling my iPhone today and getting one of each of those phones!!!

    /

  1. dagamer34

    Fresh-Faced Recruit

    Joined: Apr 2007

    +5

    Hmm.

    They are going to have to retest iOS on 4.3, as it's been recently discovered that only the Safari app gets the new JavaScript performance enhancements since they used a custom app.

  1. mytdave

    Fresh-Faced Recruit

    Joined: Aug 2000

    +4

    whatever

    So, they threw in results from iOS 4.2 running over a 3G cellular network to drag the iPhone's average down, or what?

    Apple really should fix the issues with Safari downloading one thing at a time (in serial). That is true, and it's a PITA.

  1. wrenchy

    Forum Regular

    Joined: Nov 2009

    -9

    Meaningless results.


    >>The iPhone 5 is also expected to get the same dual-core A5 chip as the iPad 2 and may get further improvements from iOS 5...


    Is Apple now trying to compete with hardware specs to compensate for a poorly coded OS/browser?

    ...Just something I would hear from iFans if the tables were turned.

  1. pairof9s

    Mac Enthusiast

    Joined: Jan 2008

    +2

    RE: Meaningless results.

    I think this is all overly exaggerated! I mean, my iPhone's web page opens 1 second slower than an Android phone. Is that REALLY impacting my day?!!

    /

  1. Sue Nommi

    Fresh-Faced Recruit

    Joined: Mar 2011

    +4

    Get some perspective

    This just appeared on this site:

    "iPhone tops JD Power ranks, widens lead over Android"
    http://www.electronista.com/articles/11/03/17/iphone.at.tops.of.jd.power.ranks.again/

    Which is more significant: overall (and trending) satisfaction...or a fixable tech spec for one particular variant of one particular Android phone?

  1. wrenchy

    Forum Regular

    Joined: Nov 2009

    -4

    @ Sue Nommi


    Who is JD Powers and Associates
    AND WHO CARES?

    That's like saying, Consumer Reports "Does not recommend" the iPhone 4

    As soon as JD P & Asses say something negative about Apple, they will be Poo Poo'd soon enough. At least from the members of this forum.

  1. bdmarsh

    Fresh-Faced Recruit

    Joined: Feb 2006

    +2

    reasonable test if you read the details

    if you read their report in full, you see that they tested both over WiFI mostly, but did test over 3G as well to see if there was a difference.

    it would be interesting to see if there was an impact to performance because the testing wasn't done with Safari itself. But generally javascript doesn't make that big of a difference.

    They mention that chrome on android loads more files at once than Safari does, and especially on full websites this makes a big difference. With more ram and processing power on their devices now, Apple may want to do the same.

Login Here

Not a member of the MacNN forums? Register now for free.

toggle

Network Headlines

toggle

Most Popular

Sponsor

Recent Reviews

Adesso Xtream S3B Bluetooth speaker

Finding a speaker purpose-built for a specific need is challenging. Even when a Bluetooth speaker can be paired with a mobile device, ...

JBL Synchros E40BT headphones

For all the different configurations of headphones on the market, it's always a tough choice for buyers to get something that is just ...

Razer Taipan mouse

The list of gaming devices is growing larger with each passing day. A large number of companies have entered the gaming input arena, a ...

Sponsor

toggle

Most Commented

 
toggle

Popular News