Apple undersells, over-delivers on iPhone 3GS speed - report

[AppleInsider]

In controlled JavaScript Web site renders, the iPhone 3GS is nearly three times as fast as the iPhone 3G and Palm Pre, and 5.5 times faster than the T-Mobile G1, according to a new study, which also reveals that the iPhone 3.0 software alone has a dramatic impact on the speed in which an iPhone 3G renders websites.

In releasing its new speed-centric update to the iPhone, Apple has boasted that the 3GS is twice as fast as its predecessor, but the company has remained mum on the handset’s actual specifications. Thanks to T-Mobile of the Netherlands, though, we know with certainty that the 3GS sports a 600MHz processor and 256MB of RAM – this compared to the 400MHz processor and 128MB of RAM on the iPhone 3G.

Based on the report released Wednesday from Medialets, a smartphone-based advertising and analytics platform, the iPhone 3GS spec bump far exceeds Apple’s “twice as fast” sales pitch in real-world tests.

Medialets’s test aimed for a fair way to compare each of the major smartphone platforms. Because they all run very different operating systems, to compare their ability to run applications would be a misnomer. However, all three platforms run browsers based on the open source WebKit standard. A MacBook running Safari was used in the test for a baseline comparison.

WebKit’s SunSpider JavaScript benchmark test was used to compare six different implementations of the WebKit browser. The test systems were:
Safari 4.0.1 on a 2.0 GHz Intel Core 2 Duo White MacBook.
Mobile Safari on the iPhone 3G with iPhone OS v2.2.1
Mobile Safari on the iPhone 3G with iPhone OS v3.0
Mobile Safari on the iPhone 3GS with iPhone OS v3.0
The “Browser” app on the T-Mobile G1 with Android OS v1.5 (Cupcake)
The “Web” app on the Palm Pre with Web OS v1.0.2
Of the mobile platforms, the iPhone 3GS came away the clear winner, besting the times of the iPhone 3G and Palm Pre by clear margins. The 3GS took an average of 16.5 seconds to render the page, while the Pre took 48.6 seconds, and the 3G running iPhone OS 3.0 took 48.7 seconds. The T-Mobile G1 took 91.1 seconds.

Another interesting outcome of the study: The iPhone 3.0 firmware has a drastic effect on the ability of the iPhone 3G to render JavaScript. The iPhone 3G running OS 2.2.1 took a whopping 132.3 seconds in the SunSpider test run by Medialets, losing clearly to even the T-Mobile G1.




A report released last week by Anandtech also showed the iPhone 3GS as the clear speed winner among all major smartphone platforms – though the end results didn’t show as drastic of an advantage for Apple’s new phone as Medialets’s study. In loading a series of popular Web sites, the Anandtech test showed the iPhone 3GS to be an average of 21 percent faster than its new WebKit-based competitor, the Palm Pre.

Of course, Web browsing only makes up one facet of the mobile phone platform. As 1up.com’s Jeremy Parish demonstrated in his game-centric 3GS review, the new iPhone’s performance boost made a huge difference in the playability of Namco Bandai's “i Love Katamari,” a game originally written for and released on the iPhone 3G. With the instant success of the iPhone 3GS in terms of sales, it is only inevitable that new games that take advantage of the speed of the new platform will begin to crop up.

For a video comparing the speed of an iPhone 3G with an iPhone 3G S when rendering websites, launching applications, and initializing games, see page 2 of AppleInsider's iPhone 3GS review.

[ View this article at AppleInsider.com ]

[Foo2]

Didn't I read somewhere that Anandtech had originally used erroneous math and that the correct speed advantage of the 3GS over the Pre was 22%, not 11%?

Anandtech updated its statistics in the second table, with a 21% advantage given to the 3GS over the Pre.
http://www.anandtech.com/gadgets/showdoc.aspx?i=3587


All this talk about iPhone killers: the 3GS is killer.

[davebroham]

3G S? I thought everyone got the memo that its now 3GS

[Quadra 610]

Well obviously. It's the iPhone, after all.

[Ireland]

Considering..

[chronster]

That's pretty cool they could speed up existing hardware through software updates. If there ever was proof that all our phones could be faster through better programming (no matter if it's windows mobile, iphone os, or android) it's this.

[Hattig]

A brilliant example of why software optimisation is still important in this day and age, when other companies just wait for hardware to improve. Apple did the double whammy.

Someone care to benchmark IE6,7 and 8 on the same MacBook hardware, to see if the iPhone 3GS can beat a desktop browser that 60%+ of the world uses?

I wonder how PocketIE does on Windows Mobile (6.0 and 6.5)? Someone, please test, if you have such a device to hand.

[wizard69]

Safari is really maturing nicely on both iPhone and Mac. Just goes to show you what good programming can do for you.

As a recently upgraded 3G owner I can say that 3.0 has extended the life of my iPhone. It should be good enough to keep me happy until next years release.


Dave

[chronster]

Originally Posted by Hattig

A brilliant example of why software optimisation is still important in this day and age, when other companies just wait for hardware to improve. Apple did the double whammy.

Someone care to benchmark IE6,7 and 8 on the same MacBook hardware, to see if the iPhone 3GS can beat a desktop browser that 60%+ of the world uses?

I wonder how PocketIE does on Windows Mobile (6.0 and 6.5)? Someone, please test, if you have such a device to hand.

I don't really know anyone who prefers pocket IE on windows mobile over opera mobile or even skyfire. The iphone might be better, but there's no point. Testing it against the newest opera 9.7 beta might be an interesting test though.

[harmsway]

My new 3G Snappy is blazing fast. the 3G is faster than using the WiFi at my work.

[solipsism]

Originally Posted by Foo2

Didn't I read somewhere that Anandtech had originally used erroneous math and that the correct speed advantage of the 3GS over the Pre was 22%, not 11%?

All this talk about iPhone killers: the 3GS is killer.

Originally Posted by davebroham

3G S? I thought everyone got the memo that its now 3GS

This comparison appears to have been written before AnandTech fixed his results, the 3GS nomenclature was changed and before the Pre updated to 1.0.3.

Originally Posted by Ireland

image: http://img151.imageshack.us/img151/6350/picture1ony.png

Considering..

image: ]http://img329.imageshack.us/img329/6946/picture2h.png

Nice find.

Originally Posted by Hattig

I wonder how PocketIE does on Windows Mobile (6.0 and 6.5)? Someone, please test, if you have such a device to hand.

For WinMo they should use Opera, like Chronster stated, as well Iris, which is the first and only mobile browser to score a 100/100 on Acid3. IT’s obviously based on WebKit.

[legend79]

To be fair; if anyone has worked with any of these artificial benchmarks before; you know as well as I do that these tests test nothing close to what webpages do. People do not write empty-loops and a lot of stupid shit like that; so I'm not sure how relevant this test really is. I guess it tells us that badly written useless code performs better on the 3GS than anything else; but last I checked; People used JS to actually do real work. Give us some REAL controlled tests already. Not synthetic!!!

[vassillios]

Safari on my 3G is Waaaaaay slower under 3.0. Tabbed view is almost unusable with 15+ secs before the page is closed after I hit the red x and then another 5+ secs for the previous page to fill the screen....and then i'm lucky if Safari doesn't crash when doing this.

[steviet02]

Originally Posted by vassillios

Safari on my 3G is Waaaaaay slower under 3.0. Tabbed view is almost unusable with 15+ secs before the page is closed after I hit the red x and then another 5+ secs for the previous page to fill the screen....and then i'm lucky if Safari doesn't crash when doing this.

I would go into the Safari Settings and clear the cache, then reboot the phone.

[solipsism]

Originally Posted by legend79

To be fair; if anyone has worked with any of these artificial benchmarks before; you know as well as I do that these tests test nothing close to what webpages do. People do not write empty-loops and a lot of stupid shit like that; so I'm not sure how relevant this test really is. I guess it tells us that badly written useless code performs better on the 3GS than anything else; but last I checked; People used JS to actually do real work. Give us some REAL controlled tests already. Not synthetic!!!

I see you point and while these tests are synthetic they are measuring many aspects of JS that are current used today or most likely will be in the near future. SquirrelFish is definitely a worthy rewrite. My biggest problem with the test is that SunSpider was developed by the WebKit team.

As for real world testing, the AnandTech article that is referenced has plenty of such tests. You can also find plenty of other sites and videos detailing the speed variances. The only problem with those is that they only give you the real world results, but they don’t help developers see what areas need more work. SunSpider is quite ideal for that.

[ai46]

Originally Posted by Hattig

A brilliant example of why software optimisation is still important in this day and age, when other companies just wait for hardware to improve. Apple did the double whammy.

Someone care to benchmark IE6,7 and 8 on the same MacBook hardware, to see if the iPhone 3GS can beat a desktop browser that 60%+ of the world uses?

I wonder how PocketIE does on Windows Mobile (6.0 and 6.5)? Someone, please test, if you have such a device to hand.

I don't have any of those products, but as an interesting aside, the iPhone 3GS gets 97/100 on the Acid3 test, while last I checked, IE6/7 gets about 21/100. While the iPhone gets a 97/100, many of the tests passed but didn't execute fast enough to not warrant a note about the speed.

[MacTel]

It is always nice to get better than spec'd performance.

Sounds like the Pre is not the iPhone killer it was made-out to be. With battery problems, dinky keyboard, and less than iPhone performance it hasn't blown away the iPhone. It at least looks nice and may give Apple some ideas on future improvements.

[DimMok]

Dudes, does anybody know of places I can get some cash for my 1G Iphone. Looking to upgrade to 3GS.....
Other than Craiglist or Ebay.....


Cause its so much snappier.

[Virgil-TB2]

Originally Posted by iPhone1982

steviet02 is right about clearing all caches and rebooting the 3G works great.

I had the same problem you were and gave it a shot. 3.0 is much quicker in all areas on my 3G. Safari page load times are MUCH quicker, mail, loading apps. Very happy with the 3.0 update so far and haven't ran into any problems or bugs yet.

While it's true that Safari shouldn't behave like that and a reboot or a cache clear will probably fix it, I have noticed more than a few pauses and glitches on my 3G iPhone running the 3.0 update as opposed to my partner's 3Gs iPhone running the same software.

Every once in a while there is a terrific pause that makes you think the thing has died, the scariest of all is that four or five times so far it has failed to register the swipe to open the home screen for up to 30 seconds or so. It always eventually opens, but there's enough time to swipe madly four or five times and just enough time for smidge of panic to set in before it does.

The software is definitely optimised for 3Gs IMO, not 3G, but I'm sure a 3.0.1 update will fix it.

[DCJ001]

Originally Posted by AppleInsider

In controlled JavaScript Web site renders, the iPhone 3GS is... 5.5 times faster than the T-Mobile G1, according to a new study...

Of the mobile platforms, the iPhone 3GS came away the clear winner, besting the times of the iPhone 3G and Palm Pre by clear margins. The 3GS took an average of 16.5 seconds to render the page, while the... T-Mobile G1 took 91.1 seconds.

Actually, the 3GS is about 5.5 times as fast as, or about 4.5 times faster than the T-Mobile G1.

[Hattig]

Originally Posted by solipsism

For WinMo they should use Opera, like Chronster stated, as well Iris, which is the first and only mobile browser to score a 100/100 on Acid3. IT’s obviously based on WebKit.

Well, seems fair to test against the shipped browser. Of course Opera can be tested as well. Never heard of Iris, but throw it in as well.

Just wanted to see the figures for academic reasons.

http://www.torchmobile.com/blog/?p=21 wrote:

Test Equipment: HTC Touch Pro, Standard ROM, connection to server via ActiveSync

Benchmark: SunSpider

Browser Results:

Access NetFront 3.5 2388.7s
Opera Mobile 9.5 (built-in) 219.0s
Pocket IE (built-in) Unable to complete
Iris Browser 1.1.2 115.0s

[jdlink]

Originally Posted by DimMok

Dudes, does anybody know of places I can get some cash for my 1G Iphone. Looking to upgrade to 3GS.....
Other than Craiglist or Ebay.....


Cause its so much snappier.

I saw mention of this site in a New York Times article. I have no experience with them myself.

http://www.gazelle.com/iphone

[teckstud]

Sounds great to me. I think I might just buy one of these (now that they are fully featured and full throttle).

[Gandalf the Semi-Coherent]

I have to agree with this analysis. Just last night, a friend of mine who purchased a 3GS, a friend with the prior 3G, and myself--with the Edge version--all did speed comparisons using Safari.

The 3GS was lightning-fast. Pages loaded almost 30 seconds faster than the 3G, and my Edge was completely left in the dust. Not only that, but apps opened twice as fast as the 3G.

Not that I needed much convincing, but I'm getting the 3GS. It's totally worth it.

--GTSC

[Virgil-TB2]

Originally Posted by iPhone1982

I have not experienced any glitches that your experiencing. I keep about 2GB free on my phone as a practice. I run 10 non Apple aps on my phone and none of them are games.

Keeping fingers crossed but only way I could be happier is if it were on a 3GS and not a 3G (always next year when my contract expires).

I usually only have 1GB or so empty on my iPhone, although that shouldn't matter. I guess it's possible that in some situations the OS uses the NAND for a cache rather than the system memory?

Either way it's a minor thing and sure to be fixed in the next minor update or firmware patch (at least based on Apple's past practices).

[shadow]

Originally Posted by Hattig

I wonder how PocketIE does on Windows Mobile (6.0 and 6.5)? Someone, please test, if you have such a device to hand.

Not all mobile browsers can run the test. On some platform it just fails, so no score to report.

Edit:
Oh, this was specifically quoted by Hattig above.

[mdriftmeyer]

Originally Posted by chronster

I don't really know anyone who prefers pocket IE on windows mobile over opera mobile or even skyfire. The iphone might be better, but there's no point. Testing it against the newest opera 9.7 beta might be an interesting test though.

Opera 10 Beta. That's the Acid 3 compliant version with their new engine.

[adnium]

Medialets’s test aimed for a fair way to compare each of the major smartphone platforms. Because they all run very different operating systems, to compare their ability to run applications would be a misnomer.

No, it would not be a misnomer, because the definition of misnomer is:

1 A wrong name or designation. LME.

S. Brett Hickton's rehearsal schedule (probably a misnomer for a process that was continuous).

2 The use of a wrong name; a misapplication of a term. M17.

B. Geldof To call the place a camp would have been a misnomer.

From the Shorter Oxford English Dictionary.

Since this is a comparative study, to compare their ability to run applications would be misrepresentative, but not a misnomer as stated in the post, since there would be no misapplication of terms.

[MacAloy]

Originally Posted by Gandalf the Semi-Coherent

I have to agree with this analysis. Just last night, a friend of mine who purchased a 3GS, a friend with the prior 3G, and myself--with the Edge version--all did speed comparisons using Safari.

The 3GS was lightning-fast. Pages loaded almost 30 seconds faster than the 3G, and my Edge was completely left in the dust. Not only that, but apps opened twice as fast as the 3G.

Not that I needed much convincing, but I'm getting the 3GS. It's totally worth it.

--GTSC

I had a 3GS the last 3 days (well, I spent all day at work with it) and noticed improvement but nowhere NEAR that much

I would say 3GS loaded 5-7 faster than 3g and Apps 2-3 seconds quicker

Improved? Yes....but nothing outragous

Wifi wise, I always thought my 3g was incredibly fast.

[IQatEdo]

Originally Posted by iPhone1982

Keeping fingers crossed but only way I could be happier is if it were on a 3GS and not a 3G (always next year when my contract expires).

The 3Gs will be my first iPhone. Next year you'll be taking delivery of a new 4Gs and we'll all be saying 'oh well, when my current contract ends'!

Comes around and goes around.

All the best.

[guinness]

Originally Posted by solipsism

For WinMo they should use Opera, like Chronster stated, as well Iris, which is the first and only mobile browser to score a 100/100 on Acid3. IT’s obviously based on WebKit.

Opera is in no shape or form, based on Webkit, they use their own engine, called Presto.

[solipsism]

Originally Posted by guinness

Opera is in no shape or form, based on Webkit, they use their own engine, called Presto.

Iris is based on WebKit, it’s the first and only current release mobile browser to score a 100/100 on Acid3. Note that a score of 100/100 does not imply passing.

[ai46]

Originally Posted by solipsism

Iris is based on WebKit, it’s the first and only current release mobile browser to score a 100/100 on Acid3. Note that a score of 100/100 does not imply passing.

For those not familiar with the Acid Tests, here is the link: <http://www.acidtests.org/>
Note that after you run the Acid3 test you can click on the "A" in Acid and get some additional comments.

[Bergermeister]

Just got my 3GS (on sale today in Japan).... WOW it's fast!

Apps open much faster, Safari loads pages faster... wow. wow. wow.

Happy camper here.