iPhone 4S Benchmarked: Obliterates competition

[jafoman]

Is this really surprising? iOS is built for one device - the iPhone. Therefore the iPhone can be optimized to take advantage of the hardware more efficiently than Android. Android is an OS built for hundreds of devices to run on. It is not optimized for a single CPU or GPU.

It's built for one device? Really?

Wow... Nobody read the actual article. AnandTech did NOT run these tests! They compiled them from other sources on the web. Read folks...

[Hardcore Til I Die]

it could be that anandtech are using stock roms and not modded ones like the ones others are using?

My Galaxy S2 scored 3096ms on Sunspider javascript benchmark.. latest stock firmware available for my phone (Android 2.3.4), dolphin browser HD. Not rooted/overclocked/etc.

[rajputwarrior]

Not really.

If the iPhone is stuck on stock, that is a limitation of the platform. If Android can be made to perform better, why is that an unfair thing?

a lot of things can be made to perform better when they are never meant to be. You think it's fair if we compare a stripped down android device that only has a browser and a couple of other services running compared to a stock iOS device.

Why would anyone want to compare an modified ROM that next to no one uses in comparison to a stock rom to iOS. that doesn't make sense... you would want to compare real world usage and what the the majority of people use, not what 1 percent (if that) of android users who eff around with their ROMs.

[Roger H. Veteran]

Well running any Android phone on AOSP builds shouldn't be counted as cheating or whatever. Google can make something great then you slap on more crap on it then of-course that will make it slower than before.

Not saying my SGS2 should be faster than anything else, I look at this like Core i5 vs Core I5 (Sandy Bridge) even at the same Ghz. Sandy Bridge version would kick pants off version 1. The A5 in the iPhone i'm sure is not the same the iPad 2 that was designed since like last summer. I'm sure they did some tweaks to the design so it's not the same CPU even though it's an A5. If that was the case then the iPad 2 would have scored the same as the iPhone 4s.

[ToneKnee]

Since I have a pretty good start with rhetorical comments in this thread… How do you explain that Safari 5.1 on my 27" iMac is so much faster than the one on my 13" MacBook Pro ? Only down to the browsers themselves, sure, sure… :laugh:

At this point, it isn’t even necessary to add that a lot of the mobile browsers out there rely on Apple’s Webkit engine… If Google modifies it to get a worse score, it’s not Apple’s problem, nor the benchmarks maker here.

Of course if you use a faster device that it can perform faster, but Browser performance is still largely down to itself and the engine behind it. If you even look at the first graph, you'll clearly see that the new safari for the iOS performs faster than the previous version because it shows the previous benchmarks. For example, I could now, go and download Firefox version 1 and compare it to the nightly build (version 10) and you will see a drastic different in Sunspider tests and every single other browser performance test. This is down to how Javascript is used, how the browser translates code to actual use in the quickest possible way without breaking standards.

If you are going to compare a system, you compare it directly with performance benchmarks like hardware, because the software changes all the time. Drivers change all the time.

You also go on to say in your point that mobile browsers out there rely on the Webkit engine. Webkit is open source, that's the reason Google use it and they've contributed largely towards it. Google haven't modified it so they can get worse scores, you are just comparing an old webkit engine to a newer one which obviously performs better. Also, how can you explain how the Honeycomb version of the browser which uses Webkit actually is in a tie with the iOS 5 version of the Safari browser?

I do have a valid argument when it comes to software benchmarks because even Anandtech even said so. This is something ICS is meant to address with mobile phones.

The one benchmark I do agree with is the actual GPU scores. This is something Apple will slap Android users with and even WP7 if it was benchmarked. This is because the hardware behind it is clearly there. As for the CPU benchmarks, this is down to the Android system being inefficient. This is something I think WP7 would score the highest on because of how well it performs as an OS.

I think we've forgotten that Android was built in the age to compete with Windows Mobile 6 and BB. Off loading the graphics to the GPU completely would give the system a big enough boost to catch up.

[TheElite]

Well, even stock ROM's are coming with far better scores than the graph shows. I think that in other tests, the 4S will not be at the top and the Galaxy S2 will not be below the iPhone 4.

[rajputwarrior]

My Galaxy S2 scored 3096ms on Sunspider javascript benchmark.. latest stock firmware available for my phone (Android 2.3.4), dolphin browser HD. Not rooted/overclocked/etc.

that's pretty good. i like that my phone will get a nice speed upgrade once i upgrade to iOS 5 (waiting for the untethered jailbreak though...). how is it with the stock browser? (as i am presuming this test was running on safari, not dolphin for iOS)

Well, even stock ROM's are coming with far better scores than the graph shows. I think that in other tests, the 4S will not be at the top and the Galaxy S2 will not be below the iPhone 4.

well according to what i quoted, it is faster then the 4S... but using a dif browser. And for your other statement, we will have to wait and see. I hope ICS is a fast and everyone thinks it'll be.

[PyX]

If the iPhone is stuck on stock, that is a limitation of the platform. If Android can be made to perform better, why is that an unfair thing?

It would be like hacking Windows’ registry on one device and then benchmark it against another. They both need the same tweaks, or if it’s not possible, the same stock installation of Windows.

[+Majesticmerc MVC]

Is this really surprising? iOS is built for one device - the iPhone. Therefore the iPhone can be optimized to take advantage of the hardware more efficiently than Android. Android is an OS built for hundreds of devices to run on. It is not optimized for a single CPU or GPU.

I think that's what is demonstrated here. I don't think Mango would perform as well either, simply for the fact that the OS has to cater to a variety of platforms and form factors, and so can only optimize for the lowest common demoninator, whereas iOS 5 has what, four or five possible hardware setups? I'm not saying this is a bad thing, it obviously works, but it would explain why it outperforms everyone else.

Honestly the biggest surprise there for me is the Galaxy S2. I hadn't realised how far ahead in the non-iPhone smartphone race they really were, and how far behind HTC are falling :o

[TheElite]

It would be like hacking Windows’ registry on one device and then benchmark it against another. They both need the same tweaks, or if it’s not possible, the same stock installation of Windows.

It isn't though.

Android is not iOS, so your analogy doesn't work here.

[bjoswald]

Congratulations Apple!

[PyX]

It isn't though.

Android is not iOS, so your analogy doesn't work here.

Ok, let me put it this way then. Who are you doing the benchmarks for? The few people out there that tweak their phone, or the majority that runs the stock version ?

[Vice]

I knew this would be a hot topic. :laugh:

[Timan Veteran]

WP Mango doesn't do too well on the browsermark benchmark, My HD7 gets slower speeds than the iphone3gs (about 20,000)

Does it really matter tho? TBH when it comes to specs, they mean ######, its how responsive a device feels that matters. Can have the fastest phone in the world, but if it still lags when swiping homescreens, or takes 5 seconds to open a browser. Means ######.

[tsupersonic]

It's built for one device? Really?

Wow... Nobody read the actual article. AnandTech did NOT run these tests! They compiled them from other sources on the web. Read folks...

Yeah iOS was designed to run on iPhone (yes there are iTouch, and iPad, but you get the point - scope of optimization Apple has to do is limited to just a handful of products). It doesn't matter where the numbers come from, benchmark are benchmarks.

[techbeck]

So, they are comparing the 4s against other phones without/no full hardware acceleration?

ICS will...and is supposed to...offer FULL hardware acceleration when GB didnt have full hardware acceleration.

So not surprised the 4s out outperforms whats currently out there for droids with similar specs. Will be interesting to see what ICS brings to the table.

And for those of you stating the android devices compared were old....I believe the Droid Bionic was just released not that long ago.

[techbeck]

that's a big IF. that's IF ICS is actually well optimized and hardware acceleration is properly done, because in honeycomb, it was not.

Hopefully they get it right with ICS as I know a lot of users are getting kinda tired of the Android OS being slow to evolve.

[KingCrimson]

I love how they tested it against all the Windows Phones :o

They compared it to phones that actually, you know... SELL units. :rofl:

[techbeck]

So what was that about hardware that was in already existing Android phones?

What about it? Hardware is the same if not better in some Android phones. I for one never said Android phones would perform better (specially when there is little to no hardware acceleration). I just said that Apple is catching up hardware wise. Thats all.

[DeathsyctheHe11]

Nice graphs.

I'm neither a Apple nor a Android fanboy, but I do hate on Android fanboys as I find them to be most obnoxious...

Every time Apple comes out with something, they never have anything nice to say. Here's some common and little bit exaggerated examples.

Apple - Android

3.5inch screen - I have 4.5inch screen that sticks out of my pockets.

Retina Display - My friend with a Samsung galaxy s II has AMOLED DISPLAY that can blow it away (but my HTC and Droid can't).

Siri - I have that like 10 years ago with vLingo.

Dual Core - My dual core overclocked 2.0Ghz can blow it away.

Notification List - My 20 year old android had that, time to play catch up Apple.

Sometimes is not all about specs and more about how well it works. Not all iPhone users use iPhones for fashion.....People buy it because it works and it works for them. If an Android works better for you, great! No need to to bash the other phone. How about congratulating them for increasing the competition?

[techbeck]

Nice graphs.

I'm neither a Apple nor a Android fanboy, but I do hate on Android fanboys as I find them to be most obnoxious...

Every time Apple comes out with something, they never have anything nice to say. Here's some common and little bit exaggerated examples.

Apple - Android

3.5inch screen - I have 4.5inch screen that sticks out of my pockets.

Retina Display - My friend with a Samsung galaxy s II has AMOLED DISPLAY that can blow it away (but my HTC and Droid can't).

Siri - I have that like 10 years ago with vLingo.

Dual Core - My dual core overclocked 2.0Ghz can blow it away.

Notification List - My 20 year old android had that, time to play catch up Apple.

Sometimes is not all about specs and more about how well it works. Not all iPhone users use iPhones for fashion.....People buy it because it works and it works for them. If an Android works better for you, great! No need to to bash the other phone. How about congratulating them for increasing the competition?

Same things can be said about any device someone is passionate about and prefers. Its only when the conversation turns ugly that its a problem. I have seen Apple and MS users make similar comments.

[+virtorio MVC]

I am quite surprised.

On paper it looks mediocre.

I guess that depends on the person reading the paper.

[PyX]

I guess that depends on the person reading the paper.

Or maybe if you transcript it wrong on the paper.

[Hardcore Til I Die]

3213.5ms using the stock browser.

[PyX]

3213.5ms using the stock browser.

You just bought a 4S? Congrats on the purchase :D

With the WiFi or with the cellular network? It still takes a few milliseconds to load theses huge scripts.

I believe they all use the WiFi to put the phones on their same basis - not all of them go up to 14.4 Mbps I guess in the list.