Preliminary Data: iPad Bests NetBook Speed at Popular Websites

| Analysis

Gomez, a company that specializes in Web performance analysis for clients, has published an initial report showing that for the top 50 most visited web pages, 74 percent of pages loaded faster on iPad than netbooks. The data is preliminary and only covered the period April 1 to 6.

The testing was restricted to users:

  • Located in the U.S. only
  • Connected via broadband connections
  • Using Wi-Fi with iPad and Netbooks

Gomez uses two metrics: 1) Raw performance, that is load time for a page. 2) Perceived performance, called render time or “above-the-fold-time.” That’s the time for all visible objects to fully render. Using these standards, Gomez found the following for the top 50 websites visited.

  1. 74% of pages loaded faster on iPad than netbooks.
  2. 62% of pages are perceived to load faster on iPad than netbooks.

Gomez also noted, however, that there are a handful of popular sites that load particularly slowly on iPad, but did not cite them by name.

The company reiterates that this is early data. Also, it should be noted that some popular websites are actively optimizing for iPad which will give the iPad a further advantage on those sites.

These results are encouraging and indicative of the potential of the iPad, its A4 processor, the Webkit engine in Safari, and what HTML5/CSS3 optimization can do. Additional, long term testing should reveal if this trend is maintained.

For more information about how Gomez, the Web performance division of Compuware, operates, see their website.

Sign Up for the Newsletter

Join the TMO Express Daily Newsletter to get the latest Mac headlines in your e-mail every weekday.

Comments

Bosco (Brad Hutchings)

Totally flawed methodology. The browser doesn’t queue up loading any Flash content. The top 50 most viewed web pages contain no Flash, right? Unreal.

studentx

Not loading Flash is a feature in my book. 90% of all Flash on the web is video and ads easily converted to HTML 5 as so many site like YouTube have already proven.

Good riddance.

daemon

studentx,

You don’t have to install flash on a netbook if you don’t want it. All netbooks are cable of having this “feature” you want.

But you know what you can’t do on the iPad? Play video on hulu.com.

ChoMomma

But you know what you can?t do on the iPad? Play video on hulu.com.

Soon…

geoduck

Disregard. Posted to the wrong story

Dean Lewis

Consider also that content providers seem to be more interested in rolling their own app or service instead of using Hulu, and Hulu may be supporting only a subscription in any app (rumor, but the New York Times isn’t exactly a total gossip rag).

I’m personally not holding my breath that Hulu will survive as is. It’s been profitable so far, but unless it figures out a way to monetize while content providers hold out for more moulah, it’s going to disappear, and that makes any app for any platform moot.

studentx

You don?t have to install flash on a netbook if you don?t want it. All netbooks are cable of having this ?feature? you want.

I don’t consider something that runs horribly slow and buggy on my computer a “feature”. Adobe had it’s chance to fix Flash on the Mac and now it’s crying a river for that failure.

As I said before, 90% of Flash is video and ads. The ads can go and the video like Hulu will be converter in-a-flash to HTML 5. Money drives the market and 100 million iPhone, Touch and iPad devices is hard to ignore, especially when those devices make up 2/3 rds of the mobile ad market.

Video is trivial to converter leave the 10% to dry up with Flash.

Bosco (Brad Hutchings)

@studentx, back when you were spermx and eggx, separated by 3 meters and a shiny brass pole, that’s what many people thought about Macs. To each their own. And to those that get in the way of us making our own choices, a bad case of blisters between your legs.

John Martellaro

The Gomez rep is looking into the concern by Bosco (first comment).  I’ll post an update when I hear back.

Bosco (Brad Hutchings)

@John. Plus which browser is used for comparison. A Flash blocker would be appropriate too on the netbook side. It just doesn’t pass the smell test though. The typical 2010 netbook is 1.6 GHz with 1 GB RAM. Users are free to install alternative browsers, such as Safari, Chrome, and Firefox. A good chunk do.

rabber

As much as I hate to do it, I need to agree with Bosco here. For the test methodology to be correct, they would need to block flash on the netbook platform and use the same browser - Safari. An interesting comparison would be the same tests on the netbook with a different browser such as IE or Firefox.

Tiger

For once, (am I really about to say this???) I agree with you Bosco.

It’s not a valid test. You can’t go comparing Apples and Lemons, I mean oranges, without equalizing your variables (and a control).

Bosco (Brad Hutchings)

What’s typically happening here is that when the browser visits a site, it needs to load 3 things. (1) The actual page. (2) The pictures on the page. (3) Embedded rich content, like Flash. Once all those are loaded, then the page will look complete. The issue is that the browser loads (1) first. Let’s simplify and assume that it completely loads (1) before starting (2) and (3) even though that’s not true. Next, the browser tries to load those things that are in (2) and (3). It might open multiple connections to the server and do some of that in parallel. It doesn’t necessary do it in any particular order though. And it might have some of those objects cached locally. The page load completion time is then dictated by how long it takes all those things to load. The dynamic (including Flash) items are typically, though not always, larger than the pictures, and if they are contending for bandwidth, they will slow down delivery of the pictures significantly when compared to if the dynamic stuff is skipped.

So what’s typical for a top 50 website (non-optimized for any platform) can be looked at with a freeway metaphor. On the iPad, the freeway is sparsely popuplated with motorcyles and small cars. The traffic doesn’t saturate the freeway and everything moves fast. On the netbook loading the Flash stuff, add in several large trucks who don’t all know about the right lane. Everybody gets slowed down. Whether faster actually works well for the conversation between end-user and website depends on what’s in those trucks. One obvious example of a top 50 site is Farmville.com or the Farmville app page on Facebook. I’m sure it loads damn quick on iPad. It’s missing, like, the whole point of the page though.

John Martellaro

It’ll be interesting to see how Gomez responds.

Jamie

It would be interesting to see those results for more reason than one - an equalized test would ironically be another black mark against Flash if it stepped up the netbook performance, don’t you think? I think it would be cool if instead of crying a river, Adobe stepped up and created a really amazing tool for HTML 5 based content. Heck, they could use different protocols and still call it Flash if they want.

Bosco (Brad Hutchings)

@Jamie, Adobe isn’t crying any river. They’re taking Flash 10.1 beta on mobile platforms soon. They’ve partnered with NVIDIA to get hardware acceleration.

And look, I know it’s probably beyond your comprehension, but whether the dynamic content is in HTML5 or Flash, the size is going to be roughly the same and the explanation I posted above will apply. But hey, when you just buy into a Steve Jobs lie, you don’t have to think before hating on things grin.

Jamie

Nah, it isn’t beyond my comprehension. I’ve worked with the web and with Adobe’s tools for 15 years, and a lot of that work has been cross-platform (it could be said that that’s nature of web work in general, and to their credit, with the recent exception of Flash-based content, Adobe has for a long time been generally very good at cross-platform solutions, in spite of their bloat). We’re all used to standards changing by now.

So far as video goes, and it’s the easiest in my opinion due to the fact that it’s already so versatile, I’d say even at this point it’s an even split with things like mpeg4, and that’s nothing new. But I think there are many factors to consider, not the least of which are the aforementioned advertising delivery and the fact that mobile consists of battery operated devices requiring protocols that will make use of resources efficiently. Additionally, new media becomes ever more interactive and less static (these are things we used to dream about with the then nascent potential of the multimedia possibilities in things like PDF back in the day); the oomph necessitated by modern apps and content makes for nit-picking on some level.

Ultimately in my experience the final decision will always rest with consumers and what they are willing to support, regardless of what kool-aid one is or isn’t drinking. It’ll be interesting to see where all of this is at in five years.

daemon

an equalized test would ironically be another black mark against Flash if it stepped up the netbook performance, don?t you think?

It’d be a bigger black mark against gomez.com for failing to provide correct benchmarking information.

Here is the meat of what gomez.com told everyone: The iPad’s A4 system-on-a-chip is more powerful than every netbook’s processor (Atom N270, Atom N450, Athlon Neo MV-40, Celeron 743, and Celeron SU2300) the most powerful of which, the Celeron SU2300, has a passmark of 914 (low powered laptops score around 1000)

And if you’re gonna claim that an ARM based processor that is marginally customized into a system-on-a-chip performs as well as a Core 2 Duo T5500 you better have raw scores available and methodology clearly detailed to back up your claims.

For reference: The Cortex-A8 ARM CPU has interger based operations on par with the Atom N450, however it’s floating-point operations were dismal. The A4 is supposed to have the Cortex-A8 and it’s possible that the iPad is offloading floating-point operations to the PowerVR SGX 535 giving the A4 vastly superior floating-point operations in comparison to other Cortex-A8 processors that have been tested.

But I’m not going to take “74 percent of pages loaded faster on iPad than netbooks” as a serious benchmark of performance when it’s just as likely that the netbooks displayed fully functional websites idententical to what you would see on a desktop compared to the neutered websites without flash and images typical of an iPad optimized website.

http://www.brightsideofnews.com/news/2010/4/7/the-coming-war-arm-versus-x86.aspx?pageid=8

daemon

I?d say even at this point it?s an even split with things like mpeg4,

You do realize that there’s a lot of different versions of mpeg4 and the only version that will work under html5 is h.264. And it’s not even clear that the majority of browsers will even implement h.264 since it’s patents have to be licensed. Ogg Theora seems to be the preferred format since there are no licenses required to implentment it’s support.

mhikl

Adobe had it?s chance to fix Flash on the Mac and now it?s crying a river for that failure.

Hey studentx,

I agree. When Flash had the biggest bag of marbles on the school yard it didn’t have to play nice so it could live by the sword, successfully. There’s a sad sense of justice seeing the bully going down as the god of comeuppance doles out a bad case of blisters eating at the bully’s bollocks. Now it’s waa waa waa from aging groupies who can’t see the future.

It took Apple, at the time a small computer company on the fringe, to overthrow the serial and parallel harlots when usb was the little wannabe on the playground. Time to move on. The future is come.

Cheers,
mhikl

Substance

I’m assuming the NetBook had Flash installed, and I believe that most of the top 50 sites use Flash in some capacity(mostly for ads, but that still counts).  And the iPad doesn’t run Flash.  Seems pretty obvious that the iPad has a pretty big advantage doesn’t it?

I also couldn’t find any information about what Web browsers were being used for the NetBook, which could also heavily skew the results.

So far, this test looks very unprofessional.

Bosco (Brad Hutchings)

Here’s a fun page that loads really quickly on iPad.

I think I’m gonna go register LoadsQuickOniPad.com. Haha.

TechInvestorNews

I work with Gomez and helped John with this post. We also assumed, as commenter Brad Hutchings did,?that the iPad-measured pages would have a smaller page weight due to the lack of Flash loading, giving?them a clear response time advantage. What we discovered when we took a closer look at the data was not so black and white.?

In 84% of?the pageviews measured on both iPads and netbooks,?the difference in page weight was less than 5%, swinging about evenly in either direction. This raises questions as to what exactly contributes to the apparent iPad performance advantage. From this preliminary data, Flash did not appear to be as big a factor as we suspected.

We are continuing to gather data and will get back to John with further insight into the Flash issue.

Bosco (Brad Hutchings)

@TechInvestorNews… It would be interesting if you could post 5 URLs that were typical in the sense of loading much faster on iPad than a Netbook.  It would also be interesting if you’d note which browser (with version) and netbook model you were testing with.

Also, on page weight… If you’re just adding up total kilobytes, that doesn’t tell half the story. You need to account for the total number of client/server interactions plus the latencies of all servers involved. If the browser is only doing, say 5 concurrent accesses, some non-Flash pieces of the page may just be waiting their turn while stuff loads from a slow third party ad server.

A fair test of rendering speed, in my mind, having given 20 seconds of thought, would be to take a web page you’re benchmarking, save it as an offline web archive, strip the main page of embedded Flash objects, host the result on a local server, and then benchmark the devices. Check server logs to ensure that no objects were pre-cached.

Log-in to comment