It appears that Apple's iPhone 4S is underclocked to 800MHz

While Apple does not reveal the exact specifications of its processing units, it appears that the new dual-core A5 chip in the iPhone 4S is underclocked to 800MHz, 200MHz less than the same chip in the iPad 2.

Despite the clocked speed, Apple has lived up to their advertising claims for the iPhone 4S.

Apple says the iPhone 4S has web browsing speeds twice as fast as the iPhone 4 and "7 times the graphic performance." The SunSpider Javascript benchmark, (run by AnandTech for browser speed) measured the 4S at 2,222ms compared to the iPhone 4's 3,921ms.

For the performance claim, Apple scored 73.1 on the GLBenchmark 2.1 compared to 11.2 for the iPhone 4.

Given the benchmark tests, the site believes the clock speed to be around 800Mhz, but as a dual-core it has significant improvements from the A4, which was underclocked to 800MHz, as well, albeit with a single core.

Written by: Andre Yoskowitz @ 11 Oct 2011 12:38

Morreale

iPhone 4S is the top performer over all those Android phones? I don't really believe that...

11.10.2011 12:49 #1

SomeBozo

Offscreen rendering? Only measures half of the equation, sure the rendering engine might be speedy, but if drawing is the bottleneck, won't matter how fast the offscreen rendering is...

11.10.2011 13:49 #2

buxtahuda

Ok. I'm about to show that I am still noobish in many ways, but what exactly is off-screen rendering? If the name is self-explanatory, I fail to see how rendering anything off the screen really matters... And I'm pretty positive there's no way an iPhone 4S is going to out-run practically anything on that list except the iPhone 4.

~*Livin' Electronicallly*~

11.10.2011 15:57 #3

i1der

Originally posted by SomeBozo: Offscreen rendering? Only measures half of the equation, sure the rendering engine might be speedy, but if drawing is the bottleneck, won't matter how fast the offscreen rendering is... I'm confused... So pure hardware with same screen still better on A5 at 800mhz (x2) against exynos 1.2Ghz (x2), maybe they should use the same software as well... Every time they compare IOS to Android is comparing a OS there very limited to another there supports much more... lets Widget, flash, cloud(yes, cloud is not out for iphone yet). maybe more focus is better for some and IOS do this great, is good to have a choice

11.10.2011 15:59 #4

dRD

Originally posted by Morreale: iPhone 4S is the top performer over all those Android phones? I don't really believe that... AFAIK, Android is built in a way that it is essentially an operating system that runs most of its processes and programs on top of a Java virtual machine, rather than as a "native app". Thus, the software performance is and will always be much worse than of those operating systems that run native apps, like WinPho, iOS and - gosh - even Symbian. Thus, it can be said that a CPU that is on paper much less powerful, but runs different OS, can produce better results than a speedier CPU that runs Android.

11.10.2011 16:16 #5

i1der

11.10.2011 16:20 #6

i1der

I just had a thought, now is time to WP7 to rise and shine...
Android is up there, and apple got stuck (my opinion), There is space between apple and others to MS start to build up to Android, Mango looks good on Demos.

11.10.2011 16:24 #7

KillerBug

You have to optimize OpenGL on android...if you use the Java-based OpenGL SDK, then you are not using the hardware OpenGL to full potential. iOS has the same problem, but since there are only a few devices, it isn't as big of an issue because you can optimize your app for all of them by doing optimization just 4 times. On android you have to either spend a ton of time optimizing for each different device, or you have to do some of the OpenGL in software so that your app runs slower, but it looks correct on all devices in spite of lack of optimization.

...or at least that is what I have been led to believe...I never tried it myself.

Personally, I couldn't care less about off-screen rendering of openGL...it has no practical use. On-screen rendering is what matters...and even then, I really don't care all that much.

IMHO, I would rather have good controls with 30FPS than terrible controls with 60FPS...and last time I checked, you still couldn't use a DS3 with an iPhone. Plus, I watch web videos a lot more than I play games on...and iOS has no flash player.

11.10.2011 16:30 #8

i1der

Originally posted by dRD: Originally posted by Morreale: iPhone 4S is the top performer over all those Android phones? I don't really believe that... AFAIK, Android is built in a way that it is essentially an operating system that runs most of its processes and programs on top of a Java virtual machine, rather than as a "native app". Thus, the software performance is and will always be much worse than of those operating systems that run native apps, like WinPho, iOS and - gosh - even Symbian. Thus, it can be said that a CPU that is on paper much less powerful, but runs different OS, can produce better results than a speedier CPU that runs Android. I'm just a bit confused about your theory, So Android, Linux and MAC runs on similar routes, and MAC vs PC(i'm pc) performance seems to run better on these what you call "essentially an operating system that runs most of its processes and programs on top of a Java virtual machine" on same lets say MACS hardware for easy compare... Seems to be a conflict here?

11.10.2011 16:32 #9

i1der

Originally posted by KillerBug: You have to optimize OpenGL on android...if you use the Java-based OpenGL SDK, then you are not using the hardware OpenGL to full potential. iOS has the same problem, but since there are only a few devices, it isn't as big of an issue because you can optimize your app for all of them by doing optimization just 4 times. On android you have to either spend a ton of time optimizing for each different device, or you have to do some of the OpenGL in software so that your app runs slower, but it looks correct on all devices in spite of lack of optimization.

...or at least that is what I have been led to believe...I never tried it myself.

Personally, I couldn't care less about off-screen rendering of openGL...it has no practical use. On-screen rendering is what matters...and even then, I really don't care all that much. Thanx, that i could relate to... the rest not so much, new device new ways to use it... would easier to have touch back like PSVita on phones, and why they keep it difficult on Bluetooth connections for controllers or other gadgets.

11.10.2011 16:38 #10

core2kid

Originally posted by Morreale: iPhone 4S is the top performer over all those Android phones? I don't really believe that... Well I can't see how 20 icons on a screen uses resources. Anyways, take an Android with a custom ROM such as CM7 and then compare it to iPhone 4S. Guarantee you the Android will kill it.

iPhone is like a Windows computer. Over time, it slows down to a crawl until you restore it. Come on, you 3G/3GS users know it's true!

11.10.2011 21:54 #11

Mrguss

iPhone 4 Vs. iPhone 4S Vs. Samsung Galaxy S II
Browser load time test....
VIDEO: http://androids.ac/65k

+4000

12.10.2011 00:58 #12

core2kid

I really don't believe in web page load times. There's a lot of factors that can affect that besides phone speed.

12.10.2011 01:02 #13

Jeffrey_P

Seems Apple fanboys do not mind spending yet more money on a so-called upgrade every year or so, That is just crazy.

Oh boy, voice recognition...

Is it really worth it? To some it must be.

My old Amiga(s) had voice recognition in the late eighties. Major problem was if you had a cold it barely would recognize the difference with voice inflection. Other then that it worked just great.

Sign a new contract sheep people.............

Oopps wrong thread, sorry.

Jeff

Cars, Guitars & Radiation.

12.10.2011 01:36 #14

dRD

Originally posted by i1der: Originally posted by dRD: Originally posted by Morreale: iPhone 4S is the top performer over all those Android phones? I don't really believe that... AFAIK, Android is built in a way that it is essentially an operating system that runs most of its processes and programs on top of a Java virtual machine, rather than as a "native app". Thus, the software performance is and will always be much worse than of those operating systems that run native apps, like WinPho, iOS and - gosh - even Symbian. Thus, it can be said that a CPU that is on paper much less powerful, but runs different OS, can produce better results than a speedier CPU that runs Android. I'm just a bit confused about your theory, So Android, Linux and MAC runs on similar routes, and MAC vs PC(i'm pc) performance seems to run better on these what you call "essentially an operating system that runs most of its processes and programs on top of a Java virtual machine" on same lets say MACS hardware for easy compare... Seems to be a conflict here? Sure, Android is Linux in its core. But the user interface and how the Android's graphical applications are run is completely different from Linux (or Mac OSX, which is a UNIX derivative). Android apps need to be run on top of Java virtualization layer. This, again, AFAIK, is required in order to ease application developers' pain to customize their apps to suit for wide variety of different CPUs, GPUs, form factors and screen sizes. But that approach has to have its toll.

Petteri Pyyny (pyyny@twitter)
Webmaster
http://AfterDawn.com/

12.10.2011 03:42 #15

Mrguss

Originally posted by core2kid: I really don't believe in web page load times. There's a lot of factors that can affect that besides phone speed. Agree:
WiFi Vs. 4G
Full Charge batt. Vs. Low change batt.
Flash On Vs. Flash Off
How many Apps. are running (Task Killer)
Gingerbread use 1 core to browse Vs iPhone 4S, Honeycomb and ICS use the 2 cores to browse........etc.

+4000

12.10.2011 07:33 #16

dali

dRD, there's an NDK for Android. Unfortunately, it's in an early phase and most of us developers are too comfortable coding still everything in Java, but critical processess should be written in C, or even other languages such as Lazarus. My personal guess is that such thing won't be happening widely enough in the near future. They'll just rely on the continuous performance boost of the processors, hardware improvements which enable faster execution of Java code, and so on. (The Nexus Prime will be a good example of this, I think).

But, anyway, it makes no sense to run a benchmark on two different platforms which is written in different languages. Specially with that language being Java on one, and C on the other. it's fair to compare iPhones this way, or different droids. But a cross-platform analysis will produce biased results like these.

"You know, it seems that quotes on the internet are becoming less and less reliable." -Abraham Lincoln.

12.10.2011 08:19 #17

dRD

Originally posted by dali: dRD, there's an NDK for Android. Alright, good to know.

Quote:But, anyway, it makes no sense to run a benchmark on two different platforms which is written in different languages. Specially with that language being Java on one, and C on the other. it's fair to compare iPhones this way, or different droids. But a cross-platform analysis will produce biased results like these. I agree wholeheartedly, there's no sense in that. Just stated the fact previously, that an average Android app (that conforms to the standards, uses Java, etc), by definition, "wastes" quite decent portion of the oomph the device itself might have, due the Android's architecture.

Anyway, as a positive sign, the fact that Android phones need bit more oomph than others, the mobile-CPU market is finally taking the steps that desktop-CPU market took back in late 1990s when we doubled the real, actual horsepower every 18 months :-) A trend that benefits all the mobile devices, not just those running Android.

Petteri Pyyny (pyyny@twitter)
Webmaster
http://AfterDawn.com/

12.10.2011 08:56 #18

dali

Agreed, dRD. And, of course, there is another important fact: When the next iPhone comes out, the 4S will be VERY underpowered when compared to contemporary Android phones, even under this kind of benchmark, because there'll be new models of many different brands with much faster processors appearing every month until then.

That is, the newest phone by Apple may be impressive at first (this isn't the case for me and most of users, though), and be massively demanded by many people when announced, but it must survive in the market for a year or two wearing the same specs... No matter the colour of the case.

And, additionally, other tech such as battery life will continue improving as well, so maybe an Android phone can be used for twice the time without needing to recharge, and, if not, well, you can always get a spare battery... Which you can't if you have an iPhone, as well as expanding or swapping flash storage, the need of the crappy iTunes and the increasing troubles to keep the phone jailbroken, etc. In short, if Apple doesn't change a lot soon, this may be its beginning of the end, because they won't be able to attract new customers, and many of the ones they have will flee to Android. Even faster now that Steve Jobs and his reality distortion field are gone.

"You know, it seems that quotes on the internet are becoming less and less reliable." -Abraham Lincoln.

12.10.2011 09:40 #19

buxtahuda

Also agreed.

Originally posted by dali: Even faster now that Steve Jobs and his reality distortion field are gone. I lul'd

12.10.2011 10:03 #20

dali

Originally posted by buxtahuda: Originally posted by dali: Even faster now that Steve Jobs and his reality distortion field are gone.
I lul'd
Hey, that term is not an invention of mine, it's Apple's.

http://en.wikipedia.org/wiki/Reality_distortion_field

:D

"You know, it seems that quotes on the internet are becoming less and less reliable." -Abraham Lincoln.

12.10.2011 11:17 #21

Mrguss

Originally posted by dali: Originally posted by buxtahuda: Originally posted by dali: Even faster now that Steve Jobs and his reality distortion field are gone.
I lul'd
Hey, that term is not an invention of mine, it's Apple's.

http://en.wikipedia.org/wiki/Reality_distortion_field

:D Well say it: "BRAVO"
RDF is use by Sales-man, Orators, Politicians, Religion Leaders, CEO's, News Reporters, Schemers, Liars or....Hey !!....why not ?....even regular people with not much knowledge about what they talking about that they believe so much on that idea and themselves, etc.

Knowledge = Power = Diff. info sources = Only time will tell us the real truth 'cos the truth is constant !

Just my 2 cents.

+4000

12.10.2011 15:31 #22

KillerBug

BTW...Android does have a native app format that does not run on top of java, but since everyone already knows java and the phones have lots of power, the developers usually just use Java. Heck, even parts of HTC sense use java...that is why it is so slow to respond versus stock android.

13.10.2011 02:21 #23

SomeBozo

Originally posted by buxtahuda: Ok. I'm about to show that I am still noobish in many ways, but what exactly is off-screen rendering? If the name is self-explanatory, I fail to see how rendering anything off the screen really matters... And I'm pretty positive there's no way an iPhone 4S is going to out-run practically anything on that list except the iPhone 4. While this is from my experience when programming on a PC, and computers were rather slow, hehe back in the day when 100Mhz was fast :) Back then developers like myself would write libraries to render stuff off screen. Essentially break the task of displaying stuff from your program into two tasks. First was to render your program's output to an off screen buffer all at once, then simply copy your off screen buffer directly to the video memory on the video card. Net difference??? Comparing times on the PC with 4.77 or 8.4Mhz, so you have some context how fast (oops slow) the machines i'm thinking of. Using the operating system without the off screen rendering might take a second or more at times, depending how much needed to be updated and displayed. In contrast the worse time usually for doing the same thing but rendering off screen and doing a memcpy call might take up to 50ms, and this was on very slow PC. I still remember 23ms was a fixed time needed to copy the memory from the off screen buffer to the video card's memory, the other 30ms was generally the time needed for the programs to generate their output.

So why that is why i question how realistic the performance of only testing the "off screen" or not. In general doing off screen rendering or not would give huge time differences. Sure, as a developer i often had to do more work to get my off screen video routines working but in the end I would normally get 10x to 100x speed gains for screen rendering.

15.10.2011 00:34 #24

KillerBug

Originally posted by SomeBozo: Originally posted by buxtahuda: Ok. I'm about to show that I am still noobish in many ways, but what exactly is off-screen rendering? If the name is self-explanatory, I fail to see how rendering anything off the screen really matters... And I'm pretty positive there's no way an iPhone 4S is going to out-run practically anything on that list except the iPhone 4. While this is from my experience when programming on a PC, and computers were rather slow, hehe back in the day when 100Mhz was fast :) Back then developers like myself would write libraries to render stuff off screen. Essentially break the task of displaying stuff from your program into two tasks. First was to render your program's output to an off screen buffer all at once, then simply copy your off screen buffer directly to the video memory on the video card. Net difference??? Comparing times on the PC with 4.77 or 8.4Mhz, so you have some context how fast (oops slow) the machines i'm thinking of. Using the operating system without the off screen rendering might take a second or more at times, depending how much needed to be updated and displayed. In contrast the worse time usually for doing the same thing but rendering off screen and doing a memcpy call might take up to 50ms, and this was on very slow PC. I still remember 23ms was a fixed time needed to copy the memory from the off screen buffer to the video card's memory, the other 30ms was generally the time needed for the programs to generate their output.

So why that is why i question how realistic the performance of only testing the "off screen" or not. In general doing off screen rendering or not would give huge time differences. Sure, as a developer i often had to do more work to get my off screen video routines working but in the end I would normally get 10x to 100x speed gains for screen rendering. Of course when you were doing all that work to squeeze a bit more performance from those old systems, there were really no dedicated video processing chips.

15.10.2011 01:59 #25

AppTrix (unverified)

Originally posted by SomeBozo: Offscreen rendering? Only measures half of the equation, sure the rendering engine might be speedy, but if drawing is the bottleneck, won't matter how fast the offscreen rendering is... oh so you think if they put it onscreen the low numbers android got are going to get any higher? ... ummmm more like all will lower.

24.10.2011 21:45 #26

Stormprobe

This site puts out misinformation. Don't use this site because you'll never know where the misinformation is going to come next.

10.12.2011 00:55 #27

AfterDawn: News