Froyo on Nexus One destroys iPhone 4 in Javascript tests

Android 2.2 Froyo and iPhone 4 with iOS 4 are now available, so the inevitable battery of comparable tests have begun. Ars Technica has taken the initiative to test Javascript performance across operating systems, and the results are not looking good on the Apple side.

SunSpider results were almost 2x slower on Safari for iPhone 4 when compared to Froyo's web browser on a Nexus One, and V8 results were over 4x better on Google's offering. They chalk this up to a JIT update and optimizations included in Froyo, but the results are even more striking when considering the handsets used to conduct the tests.

Image credit: Ars Technica

At the announcement of the iPhone 4, Apple touted its use of their A4 chip, which also powers the iPad. It appears that despite the use of a higher powered processor, Apple is still losing the speed race when it comes to Javascript rendering. Although this doesn't look great for Apple's Safari browser on iPhone, it is based on the same Webkit architecture as Android's browser, so these new Javascript runtimes should be able to slot into Apple's offering as well, should they make the effort.

Pundits have argued that frequent updates tend to cause fragmentation within the Android handset marketplace, but this is a great example of what frequent updates can offer even older handsets, when they become available.

Report a problem with article
Previous Story

Summer of Arcade games and release dates detailed

Next Story

AT&T finds slow upload speed culprit, incoming fix

68 Comments

Commenting is disabled on this article.

Ultimately due to enormous development of the underlying OS in android iOS does not stand a chance. Google is going to come at Apple with everything they have. Plus who cares, the iPhoney is a pile. Im so glad outside the USA not as many are stupid enough to follow marketing like americans.

I don't put much into javascript numbers, but what is important to notice about these tests is the page display times, especially with a device running and rendering Flash content as fast as the device that incapable of rendering the Flash content.

I'm no fan of Flash in general, but what Jobs said about Flash and mobile performance was either a lie or extremely misinformed.

(I've been running Flash on an original Droid for over a month now, and even at the default CPU clock speed, it does fine in terms of performance and doesn't eat the battery. Overclocked to 1ghz, even the heaviest Flash sites containing Video and Games work just fine.)

...it is based on the same Webkit architecture as Android's browser, so these new Javascript runtimes should be able to slot into Apple's offering as well.

the javascript engines are completely different and have little to do with Webkit. Safari plugs Apple's javascript engine (Nitro) into webkit, Android plugs in Google's (V8). this is a Nitro vs V8 contest and has little to do with mobile devices or platforms.

Funny seeing people comment about hardware acceleration and that it cannot compare to the iPhone. You know what, I am willing to bed HA is coming to the Android as well. Even if it isnt, as long as there is a demand for it it will happen. Know why? Google gives the customers what they want...and even if they didnt, there are TONS of devs out there that can make this happen.

That's awesome. Problem is while rendering speed is nice, when it comes to mobile devices the speed of the network is by far the largest differential (and considering you can find Android phones on almost any carrier, another plus for Android).

lordcanti86 said,
That's awesome. Problem is while rendering speed is nice, when it comes to mobile devices the speed of the network is by far the largest differential (and considering you can find Android phones on almost any carrier, another plus for Android).

I dont think that's true. I have a 50 Mb connection and my 3G is always at 2 Mbps. The difference between Wifi and 3G is not that big when it comes to browsing. At some point the speed is just too fast for the phone to render stuff in time.

And you seem to somehow suggest that since android phones are on any carrier, the test could have been done with a 'fast carrier for android and a slow one for the iphone4?

Julius Caro said,

I dont think that's true. I have a 50 Mb connection and my 3G is always at 2 Mbps. The difference between Wifi and 3G is not that big when it comes to browsing. At some point the speed is just too fast for the phone to render stuff in time.

And you seem to somehow suggest that since android phones are on any carrier, the test could have been done with a 'fast carrier for android and a slow one for the iphone4?


My point was that the strength and quality of your mobile connection (3G, not WiFi) will make more of a difference in your mobile browsing experience than the browser itself (unlike the desktop). It helps Android that it's available on any carrier since you can buy a phone based on which carrier has the best network where you live.

Honestly, I'm a techie guy but I really couldn't give a @#^& if a phone renders javascript slightly faster. All I really care about is features that are useful and just work when they're supposed to, like a phone's antenna/reception....

qwexor said,
Honestly, I'm a techie guy but I really couldn't give a @#^& if a phone renders javascript slightly faster. All I really care about is features that are useful and just work when they're supposed to, like a phone's antenna/reception....
Just hold it as steve says it may work for you or they will release a firmware update that will show you more bars

qwexor said,
Honestly, I'm a techie guy but I really couldn't give a @#^& if a phone renders javascript slightly faster. All I really care about is features that are useful and just work when they're supposed to, like a phone's antenna/reception....

5 second difference is "slightly" faster?

RangerLG said,

5 second difference is "slightly" faster?


LOL this is on a special script you tard. if its 0.2 or 0.1 second to load a page, you will ofcourse notice it and base your buying impulse on that?

Good for Froyo, it shows the efficiency and the productivity against the iOS4, LOL coltmann had a seniors moment lol dont worry happens to all of us. ;D

Ive owned iPhones and Androids but right now Im using the HTC Desire. I think for me, the ultimate speed test is when a web page has loaded without any notable lag or an app just works without issue. My HTC Desire does this. Im not a fan boy, I want something that works for me.

This is great news, as it shows that there is healthy competition present in the smartphone market. Apple now has some incentive to make their mobile browser better... this type of competition is what free market is all about. In the end, consumers win.

All I know, is that so far, every video that puts an iPhone up against any other phone real-world browser-performance wise, the iPhone always wins. Happened with the 3G and 3GS, and I bet it happens with the 4. Check Engadget, they've posted video speed comparisons.

King Mustard said,
All I know, is that so far, every video that puts an iPhone up against any other phone real-world browser-performance wise, the iPhone always wins. Happened with the 3G and 3GS, and I bet it happens with the 4. Check Engadget, they've posted video speed comparisons.

If you love one phone than you think its always going to win... Just shows you have no big brain,

Sebianoti said,

If you love one phone than you think its always going to win... Just shows you have no big brain,

That makes no sense. As I said, look at previous videos, the iPhone always came out on top. I see no reason why the iPhone 4 wouldn't do the same. In fact, since I wrote the previous comment, I now have video proof that Android doesn't "destroy" the iPhone 4: http://www.engadget.com/2010/0...the-browser-showdown-video/

King Mustard said,

As I said, look at previous videos, the iPhone always came out on top. I see no reason why the iPhone 4 wouldn't do the same. In fact, since I wrote the previous comment, I now have video proof that Android doesn't "destroy" the iPhone 4: http://www.engadget.com/2010/0...the-browser-showdown-video/

Without Flash the iPhone 4 wins NONE of those tests, and with it wins one. That's some nice "proof" of your claim there

IMO engadget's test is almost pointless. You got two phones with
1. roughly the same hardware
2. the same webkit html rendering engine
3. the same internet connection
doing a html page load test. So guess what, the result will be pretty much the same, as seen in the video. Who cares whether Android is half a sec faster than iOS or the other way around.

But when you're trying to load a heavy javascript based website, the difference is notable.

Edited by bestbuy, Jul 7 2010, 7:09pm :

noleafclover said,

Without Flash the iPhone 4 wins NONE of those tests, and with it wins one. That's some nice "proof" of your claim there

Yeh King Mustard = Fanboy Fail,
It goes to show how little flash slows things down.

bestbuy said,
IMO engadget's test is almost pointless. You got two phones with
1. roughly the same hardware
2. the same webkit html rendering engine
3. the same internet connection
doing a html page load test. So guess what, the result will be pretty much the same, as seen in the video. Who cares whether Android is half a sec faster than iOS or the other way around.

But when you're trying to load a heavy javascript based website, the difference is notable.

So to you the test that's worthwhile is the rare pages you visit that have enough javascript to make a difference, and not the every day browsing that most people will be doing? That makes sense.

King Mustard said,
All I know, is that so far, every video that puts an iPhone up against any other phone real-world browser-performance wise, the iPhone always wins. Happened with the 3G and 3GS, and I bet it happens with the 4. Check Engadget, they've posted video speed comparisons.

Dont you realize that when the 3g and 3gs were released that there was basically no competition with the iPhone? So of course they will come out on top. Now there is competition and Apple seems to be lagging and of course producing faulty products.

King Mustard said,
All I know, is that so far, every video that puts an iPhone up against any other phone real-world browser-performance wise, the iPhone always wins. Happened with the 3G and 3GS, and I bet it happens with the 4. Check Engadget, they've posted video speed comparisons.

Engadget aka AppleFanboigadget? Engadget has 0 credibility anymore with anything to do with Apple after their ridiculously lop-sided iPhone 4 review complete with antenna issues. But hey, if it makes you sleep better *thinking* you have the baddest phone on the planet, who cares.

Betaz said,

So to you the test that's worthwhile is the rare pages you visit that have enough javascript to make a difference, and not the every day browsing that most people will be doing? That makes sense.

since the "every day browrsing" performance is pretty much the same (in fact i think the bottleneck is more on the internet connection) , you gotta look at some other things that do make a difference. Just like Flash, do you use it on your phone on a dialy basis? no. But when you need it it's there. Same goes for JS performance, it's there when you need it.

bestbuy said,

since the "every day browrsing" performance is pretty much the same (in fact i think the bottleneck is more on the internet connection) , you gotta look at some other things that do make a difference. Just like Flash, do you use it on your phone on a dialy basis? no. But when you need it it's there. Same goes for JS performance, it's there when you need it.

Oh, I agree, and their javascript IS much faster, there's no argument about it. It's Android fans that make a huge deal out of it like it's the one thing that makes their phone amazing and the iPhone suck, however, the iPhone doesn't load pages any slower in the grand scheme. Javascript benchmarking is simple numbers benchmarking and nothing more. Every fraction of a second more is obviously a good thing, but the large majority of websites use so little javascript per page load that it boils down to nearly nothing in real life scenarios. It's a good thumbs up to Google, but it's nothing to even consider when choosing which phone to buy.

Betaz said,

Oh, I agree, and their javascript IS much faster, there's no argument about it. It's Android fans that make a huge deal out of it like it's the one thing that makes their phone amazing and the iPhone suck, however, the iPhone doesn't load pages any slower in the grand scheme. Javascript benchmarking is simple numbers benchmarking and nothing more. Every fraction of a second more is obviously a good thing, but the large majority of websites use so little javascript per page load that it boils down to nearly nothing in real life scenarios. It's a good thumbs up to Google, but it's nothing to even consider when choosing which phone to buy.


However, Apple is really behind the ball on this, considering they are pushing HTML5/JS for rich content instead of flash. Converting flash functionality to equivalent HTML5 is going to take a lot of JS, so JS performance is going to be very important in the near future, especially on platforms not supporting flash (iOS).
If Android (on slower? hardware) has better JS performance than iOS, and the ability to run flash, it certainly makes it a more attractive platform for rich web browsing than the iphone.

King Mustard said,
All I know, is that so far, every video that puts an iPhone up against any other phone real-world browser-performance wise, the iPhone always wins. Happened with the 3G and 3GS, and I bet it happens with the 4. Check Engadget, they've posted video speed comparisons.

Really, like in tests like this?
http://www.youtube.com/watch?v...ehV7KhU&feature=related

Most ZuneHD vs iPod/iPhone video comparisons are from when the ZuneHD was launched before the browser update a couple of months later. All the video tests you find between any iPod touch or iPhone and the ZuneHD after the browser update on the ZuneHD shows the ZuneHD to be faster. (Also remember this is a variant of fairly old IE code running on the ZuneHD's WinCE OS.)

The iPhone is nice, but it is NOT always the best or fastest for everything. Get out of that mindset or Apple will be selling you iCrap with rolls of paper and you will consider buying it.

The results are certainly noteworthy, but it's worth noting this isn't exactly an apples-to-apples comparison, to the extent that rather than comparing "raw" performance of the hardware, you have the (noted) difference in the actual browsers, despite their being based on the WebKit engine. Because of this software difference it speaks to Google's "better"/speedier/whatever programming, but it doesn't directly compare the Nexus One hardware to that of the iPhone 4.

Aeden said,
The results are certainly noteworthy, but it's worth noting this isn't exactly an apples-to-apples comparison, to the extent that rather than comparing "raw" performance of the hardware, you have the (noted) difference in the actual browsers, despite their being based on the WebKit engine. Because of this software difference it speaks to Google's "better"/speedier/whatever programming, but it doesn't directly compare the Nexus One hardware to that of the iPhone 4.

you seems to miss the hole point of that comparison...

Aeden said,
The results are certainly noteworthy, but it's worth noting this isn't exactly an apples-to-apples comparison, to the extent that rather than comparing "raw" performance of the hardware, you have the (noted) difference in the actual browsers, despite their being based on the WebKit engine. Because of this software difference it speaks to Google's "better"/speedier/whatever programming, but it doesn't directly compare the Nexus One hardware to that of the iPhone 4.

This is most certainly "apples to apples". There is no mention of hardware comparison here, this is a comparison in each phone's ability to process JavaScript via their stock browser.

Aeden said,
The results are certainly noteworthy, but it's worth noting this isn't exactly an apples-to-apples comparison

I know, because this is an Apple to Google comparison.

ccoltmanm said,
5000 milliseconds! That's one half of one second!

Actually, 5000ms is 5 seconds, which is much more significant than half of a second.

ccoltmanm said,
5000 milliseconds! That's one half of one second!

I agree....Some would see that as a hell of a lot of time depending on the environment that 0.5 second time frame is being measured and what on.

cmcasey79 said,

Actually, 5000ms is 5 seconds, which is much more significant than half of a second.

Oh yeah...blond moment whilst reading ccoltmanm's comment!

androidking said,

I agree....Some would see that as a hell of a lot of time depending on the environment that 0.5 second time frame is being measured and what on.


As stated above, 5000ms is 5 seconds. 5 seconds is a long time for anything on the web...

JonathanMarston said,

As stated above, 5000ms is 5 seconds. 5 seconds is a long time for anything on the web...

As stated above.....Oh yeah...blond moment whilst reading ccoltmanm's comment!

JonathanMarston said,

As stated above, 5000ms is 5 seconds. 5 seconds is a long time for anything on the web...

In school for web design, we were told the average amount of time people are willing to wait for a page to load is 7 seconds. Then they lose interest and look elsewhere.

I don't know what this test is as I've been out of the game for a while but it's a test. At least I don't know how big it is or how long it should take but when you have one take 57 seconds and another 109, that's a huge difference.

dogmai said,

In school for web design, we were told the average amount of time people are willing to wait for a page to load is 7 seconds. Then they lose interest and look elsewhere.

I don't know what this test is as I've been out of the game for a while but it's a test. At least I don't know how big it is or how long it should take but when you have one take 57 seconds and another 109, that's a huge difference.

5.7 and 10.9, not 57 and 109.. 1000ms = 1s.

dogmai said,

In school for web design, we were told the average amount of time people are willing to wait for a page to load is 7 seconds. Then they lose interest and look elsewhere.

I don't know what this test is as I've been out of the game for a while but it's a test. At least I don't know how big it is or how long it should take but when you have one take 57 seconds and another 109, that's a huge difference.

You're not understanding the benchmark. Nothing was tested to see how long a page takes to load. Head over to Engadget to look at their browser comparison for that. This is just a few javascript tests to see how long it takes for a browser to run through a variety of javascript functions. In reality, most pages have very little javascript comparatively that has to load on page load, so this has only a small fraction of a difference in load time.

All this is doing is benchmarking the javascript engine. Android's is much faster than the iPhone's, however, the iPhone 4 still loads pages at almost exactly the same speed.

King Mustard said,
Some website elements are hardware-rendered in iOS, unlike in Android. That makes a big difference.

Ok, so why are the elements hardware rendered in iOS, but not in Android? What are the pros and cons?

serious said,

Ok, so why are the elements hardware rendered in iOS, but not in Android? What are the pros and cons?
Because it will be faster. Rendering in hardware is faster than in software.

pickypg said,
Because it will be faster. Rendering in hardware is faster than in software.

Faster, as in 2x slower than Android?
That's not faster.

King Mustard said,
Some website elements are hardware-rendered in iOS, unlike in Android. That makes a big difference.

With the JIT compiler included in Froyo, this will make very little difference.

Alastyr said,

Faster, as in 2x slower than Android?
That's not faster.

these differences dont say ****, its just marketing

Chrome is the fastest on paper, yet seeing by eye... FF is equally as fast in loading pages (sometimes a bit slower, not much) and IE8 is allot faster, IE8-64bit is even faster then the 32bit...

tests dont say a darn thing...

King Mustard said,
Some website elements are hardware-rendered in iOS, unlike in Android. That makes a big difference.

You forgot: ...Apple claims.

Apple also claims that they are hardware assisting HTML5, and they very well may be doing this, but Safari is still slower at showing the final content or dealing with dynamic content than browsers like Opera or even IE8.

Real hardware accelerated 'rendering' like in IE9 is 100x faster than Safari or page rendering on iOS's Safari.

I do agree with you that javascript is NOT a complete test of browser performance. IE8 shows horrible javascript numbers, yet keeps up with Chrome and Firefox on general surfing because it does 'render' some things faster and there is more to browser performance that just the scripts, there is the parsing, the data retrieval, and also dealing with dynamic content once the page initially loads.

However, Apple isn't doing any of these areas well, and even in the one or two areas they do fight for the numbers they are still behind Opera and Chrome and in the mobile world behind the Android browser and even behind the IE browser on devices like the ZuneHD that is running a very old IE engine, yet loads pages several times faster than the iPhone 3gs or the iPod Touch.

Apple is essentially 'hardware assisting' like they did GUI animations with a kludge shoved through SSE because they didn't standardize the GPU on Macs to meet a minimum level of performance.

What Apple is showing today with HTML5 and their 'hardware assisted' rendering is a full generation even Firefox's Direct2D rendering on Vista and Win7 and two generations behind IE9 GPU assisted and Direct2D rendering.

Apple needs to wake up and find a way to get a faster path to the display on both iOS and OS X if they will ever be able to compete as even a Web browsing device/computer once HTML5 becomes the status quo and site designers are baselining on Firefox or IE9.

Google also needs to do this as well with Chrome and the Android browser and so far neither Google or Apple or showing signs of 'getting it' and creating an engine that drops as much HTML as possible to hardware and not just Javascript.

Microsoft is set to rule if Google or Firefox or Apple doesn't wake up and FAST.

IE9 compared to the Direct2D hardware version of Firefox is also a massive difference. On the HTML5 canvas tests Firefox holds its own, but if you run any of the tests that are based on the other aspects of HTML5, Firefox is 10x or more slower than IE9 because Firefox is JUST using the Direct2D as the final rendering instead of GDI. In contrast IE9 is shoving every bit they can of the page content through hardware, creating an almost 'compiled' level of performance with HTML/CSS and other elements of what is known as HTML5. (Many browsers are using a JIT for javascript, but none are applying this concept to the other aspects and content of the web page.)

Shadowzz said,
these differences dont say ****, its just marketing

Chrome is the fastest on paper, yet seeing by eye... FF is equally as fast in loading pages (sometimes a bit slower, not much) and IE8 is allot faster, IE8-64bit is even faster then the 32bit...

tests dont say a darn thing...

I second that - Microsoft did a video not too long ago which measured the different steps each take when rendering a page, Javascript makes up a very small amount of that process. I would sooner see browser vendors spend more time on security, reliability and rendering rather than jerking off over the supposed merits of their particular javascript engine.