iPhone 3GS Tops T-Mobile G1 and Palm Pre in JavaScript Benchmarks

[MacRumors]

Medialets today released the results of JavaScript benchmark tests performed using the SunSpider test suite on the iPhone 3G and 3GS, the T-Mobile G1 running Android, and the Palm Pre running webOS. The use of the SunSpider suite for benchmarking studies enables cross-platform comparisons among the devices, as they all utilize WebKit-based Web browsers. In the Medialets study, the iPhone 3GS outperformed the Palm Pre by a factor of three and the T-Mobile G1 by over a factor of five.

Notably, Medialets also tested the iPhone 3G running both iPhone OS 2.2.1 and OS 3.0, revealing that iPhone OS 3.0 provides a nearly three-fold improvement in performance over OS 2.2.1 running on the same hardware. The shift to the iPhone 3GS increases performance a further three-fold. All results were compared to the performance of a 2 GHz Core 2 Duo White MacBook, demonstrating JavaScript performance on the iPhone 3GS only 12X that of the MacBook, a remarkable statistic for a smartphone.




Medialets cautions readers that the JavaScript benchmark results do not necessarily indicate that one device or operating system is "better" than another, and that each operating system may be more well-suited for certain tasks than the others.

Article Link: iPhone 3GS Tops T-Mobile G1 and Palm Pre in JavaScript Benchmarks

[mobi]

Real screamer!

I want to test myself real-world side by side...

[mkrishnan]

The findings related to the 2.2.x -> 3.0 jump on old hardware are very happy, too. Impressive that the 3G can keep up with the Pre on this particular metric with the new firmware. The rendering speed changes seem evident to me on my EDGE iPhone also.

[HyperZboy]

Why didn't they test the original iPhone as well?

I'd love to see how the improvement is even on EDGE with the 3.0 software.

I'd laugh my ass off if it beat any other 3G phone just from the Javascript performance, but I guess that's unlikely.

Would still like to know though just for fun.

[elppa]

Originally Posted by HyperZboy

Why didn't they test the original iPhone as well?

I'd love to see how the improvement is even on EDGE with the 3.0 software.

Javascript is interpreted on the client side, which means that the network the script travelled over to arrive at the client is irrelevant.

What this does prove is Palm's bravado about “OS X being too big, bloated and slow for mobile devices” is mostly nonsense as each operating system may be more well-suited for certain tasks than the others.

[LukeHarrison]

Originally Posted by HyperZboy

Why didn't they test the original iPhone as well?

I'd love to see how the improvement is even on EDGE with the 3.0 software.

I'd laugh my ass off if it beat any other 3G phone just from the Javascript performance, but I guess that's unlikely.

Would still like to know though just for fun.

I would imagine that as the iPhone 2G has the same internals as the 3G, that the benchmark applies to it also.

Hopefully.

[mkrishnan]

Originally Posted by HyperZboy

I'd love to see how the improvement is even on EDGE with the 3.0 software.

I'd laugh my ass off if it beat any other 3G phone just from the Javascript performance, but I guess that's unlikely.

From what I can tell of this test, this is a test of interpreter speed -- that is, it's not dependent on transmission speed. So the EDGE/3G aspect doesn't really play a role. The original iPhone would be slower because of other things about it that are (modestly, and somewhat more pronounced in things like games) slower than the 3G, as I understand it. But basically, in terms of what this test tests, the performance of the original iPhone should actually be fairly close to the 3G on 3.0.

[iphones4evry1]

So what exactly was measured ... the rendering of Java?
In other words, was it the time it took for the processor to render the Java after it had been downloaded?

"JavaScript execution in WebKit—the open source project that, in varying degrees, powers web browsing technology for these three disparate operating systems. - web rendering technology"

"this benchmark tests the core JavaScript language only, not the DOM or other browser APIs. It is designed to compare different versions of the same browser, and different browsers to each other."

"Network speed and latency have no effect on the results of the test."

Link to original article: http://www.medialets.com/blog/2009/0...-apple-claims/

.

[DipDog3]

That is still quite impressive!

[mkrishnan]

Originally Posted by iphones4evry1

In other words, was it the time it took for the processor to render the Java after it had been downloaded?

Not Java. Javascript. They're two almost completely different things. Javascript is a kind of code that allows web pages to have "dynamic" content -- it underlies a lot of how web pages show different kinds of information and get customized based on user clicks and so forth. It's a pretty major part of the modern internet. Javascript is sent as text code to the browser and then rendered on the computer or phone. So rendering Javascript is a pretty significant part of how fast a website displays, aside from how fast the internet connection works.

So Jscript is one reason why, when you put your iPhone and your notebook computer on the same WiFi network, your iPhone still takes much longer to render a page than your computer does.

It's also the major reason why Safari and Chrome are said to be so much faster than other web browsers -- recently there's been a lot of progress in how to render Javascript, and most of the speed difference between Safari and Chrome and... Firefox, Opera, and IE, is because of javascript rendering speed.

[aristotle]

Originally Posted by iphones4evry1

So what exactly was measured ... the rendering of Java?
In other words, was it the time it took for the processor to render the Java after it had been downloaded?

"JavaScript execution in WebKit—the open source project that, in varying degrees, powers web browsing technology for these three disparate operating systems. - web rendering technology"

Link to original article: http://www.medialets.com/blog/2009/0...-apple-claims/

.

Can't you read? It says quite plainly in what you quoted that these tests measured the time to execute (run) the javascript code on the test pages. Java has nothing to do with it as this is Javascript, not Java.

[iphones4evry1]

Originally Posted by mkrishnan

Not Java. Javascript. They're two almost completely different things. Javascript is a kind of code that allows web pages to have "dynamic" content -- it underlies a lot of how web pages show different kinds of information and get customized based on user clicks and so forth. It's a pretty major part of the modern internet. Javascript is sent as text code to the browser and then rendered on the computer or phone. So rendering Javascript is a pretty significant part of how fast a website displays, aside from how fast the internet connection works.

So Jscript is one reason why, when you put your iPhone and your notebook computer on the same WiFi network, your iPhone still takes much longer to render a page than your computer does.

It's also the major reason why Safari and Chrome are said to be so much faster than other web browsers -- recently there's been a lot of progress in how to render Javascript, and most of the speed difference between Safari and Chrome and... Firefox, Opera, and IE, is because of javascript rendering speed.

Thank you for explaining.

[Mr Maui]

There are those out there who will still complain that despite these types of results from independent testers, that the G3S is not fast enough, and will further complain that although it is faster and offers more/improved features (for the same price) from the 3G, that it is STILL a MINOR upgrade. Hopefully they'll all find themselves posting on PreRumors.com (way popular) or G1Rumors.com (way, Way, WAY popular).

[Daniel0418]

I am really starting to not trust these benchmarks. At my house on wifi and with 3G my dads pre always loads faster than my 3GS and I have 5 bars he only has 3... Also on c.net the pre won aswell in sending pictures and rendering pages. And my sisters G1 is slower than my 3GS but not by that much... I am starting to sense a huge bias. Oh well never can trust the media.

[windywoo]

If my phone cost twice as much as the other phones I would expect it to be better in Java benchmarks.

[Goona]

Originally Posted by Daniel0418

I am really starting to not trust these benchmarks. At my house on wifi and with 3G my dads pre always loads faster than my 3GS and I have 5 bars he only has 3... Also on c.net the pre won aswell in sending pictures and rendering pages. And my sisters G1 is slower than my 3GS but not by that much... I am starting to sense a huge bias. Oh well never can trust the media.

Right there is a huge media bias against the Pre after they hyped it for the past 6 months.

[Goona]

Originally Posted by windywoo

If my phone cost twice as much as the other phones I would expect it to be better in Java benchmarks.

Which phone is that?

[sellitom]

iPhone ftw.

[iPhoneNYC]

I find it very interesting of the speed bump from just the new software. I'm sure that will make the folks waiting to upgrade happy.

[Lesser Evets]

Wowzers.

What will it be in 2 or 3 years when Apple has fully fleshed out their portable phone/internet tech? 2x baseline of a low end notebook? Technology is going to be leaping by the 20-teens.

[Prof.]

Is anyone actually shocked by this? Of course the iPhone is faster.

Superior Software + Superior Hardware = Superior Product.

[JBaker122586]

Originally Posted by Daniel0418

I am really starting to not trust these benchmarks. At my house on wifi and with 3G my dads pre always loads faster than my 3GS and I have 5 bars he only has 3... Also on c.net the pre won aswell in sending pictures and rendering pages. And my sisters G1 is slower than my 3GS but not by that much... I am starting to sense a huge bias. Oh well never can trust the media.

First of all, this is about Javascript interpretation within the web browser.
Nothing more, nothing less. These are with benchmark tests I believe provided by Sun, so there's absolutely no bias involved.
Whether Sprint's 3G is faster than AT&T or has a stronger signal at your house or the CNet office has absolutely no bearing on the JS rendering ability of the phone.

[GoneFishing]

No competition here. Obviously the White Macbook finished first in the group. Guess I'm getting a macbook then.

[JBaker122586]

Originally Posted by windywoo

If my phone cost twice as much as the other phones I would expect it to be better in Java benchmarks.

They're not comparing it to the Razr, they're comparing it to the Pre (you know, the $300-$100MIR phone with 8GB of storage), and the G1 (the $150 phone with 256MB of built-in storage).

[InkMaster]

Originally Posted by windywoo

If my phone cost twice as much as the other phones I would expect it to be better in Java benchmarks.

Then going by that theory, shouldn't some of the high end Nokia phones blow away anything else? ...in theory at least