PDA

View Full Version : What are Samsung doing with all that power? iPhone 5 - 1/2 the specs as S3 and faster




gwelmarten
Sep 16, 2012, 03:37 PM
Hi
I thought I'd open this thread for people to talk about the difference in Geekbench scores between the iPhone 5 and the Samsung S3. There's a rather odd thing going on.



So, the iPhone 5.
Has: Dual core 1.02GHz CPU and 1GB of RAM.
Apple do not heavily advertise the specs, as it doesn't matter to consumers.

and the Samsung S3
Has: Quad Core 1.4GHz CPU and 2GB of RAM.
Samsung advertise this heavily, making out to the consumer that the extra power will make their life better.

BUT...
the Geek Bench scores show something different. This test is renowed for being the best test of the performance of a device. The scores are as follows:
iPhone 5: 1601
Samsung S3: 1560

What are Samsung doing with all this extra power? Leaking memory and sending your personal information to advertisers? How can they sleep at night advertising that a phone is more powerful when it's just slower? Do they just think "Hmm - phone seems to be running a little slowly... maybe we should consult our software - look for leaks and more efficient ways of doing things. Naaaa, lets just throw more POWER at it and advertise it!!!"



I'd be interested to hear other people's opinions on this. Any other thoughts?

Sam



jabingla2810
Sep 16, 2012, 03:39 PM
Samsung throw faster processors and more RAM into a phone to make it run smoothly.

Apple optimise software and hardware to work well together.

617aircav
Sep 16, 2012, 03:39 PM
Who cares enjoy your phone

matttye
Sep 16, 2012, 03:40 PM
Full multitasking, widgets, smart stay, numerous gestures, application voice commands,... These are just a few of the things constantly running on the S3 that aren't on the iPhone.

gwelmarten
Sep 16, 2012, 03:41 PM
Full multitasking, widgets, smart stay, numerous gestures, application voice commands,... These are just a few of the things constantly running on the S3 that aren't on the iPhone.

Yes, but this seriously results in a degraded battery life.

jav6454
Sep 16, 2012, 03:41 PM
Ok, this should be in the main thread, but what the hell. I'll bite. Android isn't a resource friendly OS. Much like Windows manufacturers advertise heavily on specs where as Apple doesn't need that much. Reason?

Android is a general OS for all types of phones. So it is in a sense bloated and has to account for several types of configurations as possible.

iPhone OS, is custom tailored to just one set of hardware. Hell, iPhone OS 6.0 dropped the support for iPhone and iPhone 3G in its entirety. Only support for the iPhone 3GS, iPhone 4 and iPhone 4S remain. Apple continuously optimizes iPhone OS to make it as fast and performance efficient with as little as possible.

Result? S3 is beaten in Geekbench scores.

TSX
Sep 16, 2012, 03:42 PM
I thought the us version of the s3 was dual core and the international was quad?

zachnelson
Sep 16, 2012, 03:42 PM
Android apps and the framework are written in Java, which requires significantly more processing power than iOS apps and the framework require since they are written in.. C++ I believe.

Oppressed
Sep 16, 2012, 03:43 PM
Specs don't matter by themselves. Apple makes the software to take full and complete advantage of every piece of silicon in their devices, and this proves it. Meanwhile companies, like Samsung,have no option but to shove more specs into their phones and hope it runs a little better.

I can see how this may be a hard concept to get for some die hard Fandroids who came from the PC spec game of the 2000's.

jrlepage
Sep 16, 2012, 03:43 PM
Samsung throw faster processors and more RAM into a phone to make it run smoothly.

Apple optimise software and hardware to work well together.

^this

It takes better specs to run a clunky OS.

matttye
Sep 16, 2012, 03:44 PM
Yes, but this seriously results in a degraded battery life.

That wasn't part of your question, but I'll entertain it anyway. Battery life is adequate.

358012

358013

gwelmarten
Sep 16, 2012, 03:46 PM
Ok, this should be in the main thread, but what the hell. I'll bite. Android isn't a resource friendly OS. Much like Windows manufacturers advertise heavily on specs where as Apple doesn't need that much. Reason?

Android is a general OS for all types of phones. So it is in a sense bloated and has to account for several types of configurations as possible.

iPhone OS, is custom tailored to just one set of hardware. Hell, iPhone OS 6.0 dropped the support for iPhone and iPhone 3G in its entirety. Only support for the iPhone 3GS, iPhone 4 and iPhone 4S remain. Apple continuously optimizes iPhone OS to make it as fast and performance efficient with as little as possible.

Result? S3 is beaten in Geekbench scores.

This compares Android to Windows (a comparison I had already made). But in this case, there are still massive optimisations that can be made without loosing compatibility with sets of hardware. With a comparatively small effort on Samsung's part (and other Android phone manufacturers) they could modify Android to be 75% more efficient - lots of that code is easily configurable to your platform.

----------

That wasn't part of your question, but I'll entertain it anyway. Battery life is adequate.

358012

358013

Ok - point made with the battery. How do they do that compared to the iPhone? Apple always BS'ing about how great their battery is, and comparatively to my other Samsung stuff, my iPhone battery is crap.

vistadude
Sep 16, 2012, 03:50 PM
How can you say the iphone 5 is faster without having used one? The galaxy S3 can show videos and widgets on the homescreen. The iphone cannot. The S3 can let you talk on the phone and use data, send text messages, run gps navigation, simultaneously, on any LTE network. The iphone cannot do this.

Seems iphone can't even do basic features?

zachnelson
Sep 16, 2012, 03:54 PM
This compares Android to Windows (a comparison I had already made). But in this case, there are still massive optimisations that can be made without loosing compatibility with sets of hardware. With a comparatively small effort on Samsung's part (and other Android phone manufacturers) they could modify Android to be 75% more efficient - lots of that code is easily configurable to your platform.

Read below please,

Android apps and the framework are written in Java, which requires significantly more processing power than iOS apps and the framework require since they are written in.. C++ I believe.

Samsung can not do anything to modify Android to be more efficient. Android is written in Java, an inherently un-efficient and processor heavy coding language.

----------

How can you say the iphone 5 is faster without having used one? The galaxy S3 can show videos and widgets on the homescreen. The iphone cannot. The S3 can let you talk on the phone and use data, send text messages, run gps navigation, simultaneously, on any LTE network. The iphone cannot do this.

Seems iphone can't even do basic features?

A bench mark was just released that blows the S3 out of the water with pure processing power.

The ability to have power draining widgets on a homescreen, using data + voice at the same time or other silly things fandroids use to buff their phones up are not related to the speed of the device.

gwelmarten
Sep 16, 2012, 04:01 PM
How can you say the iphone 5 is faster without having used one? The galaxy S3 can show videos and widgets on the homescreen. The iphone cannot. The S3 can let you talk on the phone and use data, send text messages, run gps navigation, simultaneously, on any LTE network. The iphone cannot do this.

Seems iphone can't even do basic features?

How can you say the iphone 5 is faster without having used one?
The Geek Bench score shows so, and Geek Bench is widely accepted to be right.

The galaxy S3 can show videos and widgets on the homescreen.
Apple choose not to offer this as a feature. And thank god. Why would I want a video of my cat in repeat all day every day?

The S3 can let you talk on the phone and use data, send text messages, run gps navigation, simultaneously, on any LTE network. iPhone now does this. It was only the CDMA iPhone 4 that could not do this, due to a network restriction.

vistadude
Sep 16, 2012, 04:09 PM
Benchmarks might test raw processing power, but they don't show which is faster. Which phone can boot up faster, which phone can start a skype call faster, which phone lets you make a phone call faster? These are things that actually show if a phone is faster.

I guess you like to run one program at a time, just like a true apple fanboy. In the windows world, when a person opens up a youtube video, you can have 10 other screens open, check email, do instant messaging, and play a game with other friends. I'd sure like to do this on a phone, especially when a video is buffering.

The CDMA iphone 5 won't let you talk on the phone and use data, receive text message, receive notifications, and skype. But the S3 will!


How can you say the iphone 5 is faster without having used one?
The Geek Bench score shows so, and Geek Bench is widely accepted to be right.

The galaxy S3 can show videos and widgets on the homescreen.
Apple choose not to offer this as a feature. And thank god. Why would I want a video of my cat in repeat all day every day?

The S3 can let you talk on the phone and use data, send text messages, run gps navigation, simultaneously, on any LTE network. iPhone now does this. It was only the CDMA iPhone 4 that could not do this, due to a network restriction.

McCool71
Sep 16, 2012, 04:21 PM
Hi
BUT...
the Geek Bench scores show something different. This test is renowed for being the best test of the performance of a device. The scores are as follows:
iPhone 5: 1601
Samsung S3: 1560


Really? This is what I get when I search:

Iphone 5 (1601 points):
http://browser.primatelabs.com/geekbench2/1030202

Galaxy S3 4-core, 1GB (Non-US version) (1847 points):
http://browser.primatelabs.com/geekbench2/998950

Galaxy S3 2-core, 2GB (US version) (1683 points):
http://browser.primatelabs.com/geekbench2/884649

That said, pure processing power isn't necessarily a good measuring stick for how a cell phone performs in day-to-day activities.

Technarchy
Sep 16, 2012, 04:22 PM
As any person that has used android for awhile can tell you the OS requires heaps of high powered silicon to run well. It's a shoddy mobile OS.

It's not efficient and it doesn't scale worth a damn, so no, a jellybean ROM wont solve your problems.

Either way, I don't like Geekbench. I prefer seeing Anandtech's in depth look, but in the end, it really doesn't matter.

I'm never buying another android device, and my iPhone 5 should be here on Friday.

G51989
Sep 16, 2012, 04:23 PM
Yes, but this seriously results in a degraded battery life.

Not really

matttye
Sep 16, 2012, 04:25 PM
Ok - point made with the battery. How do they do that compared to the iPhone? Apple always BS'ing about how great their battery is, and comparatively to my other Samsung stuff, my iPhone battery is crap.

Galaxy S3 has a much larger battery than the 4S, not sure about the 5. My battery would probably be about 10% lower than it is if I hadn't been on WiFi all day.

It's acceptable though, it easily lasts from when I wake up to when I go to bed with what I would consider to be heavy use.

ChazUK
Sep 16, 2012, 04:27 PM
I've bench my Exynos powered S3 in Geekbench alongside my Tegra3 powered Nexus 7 and the results are as follows:

Nexus 7 = 1554
Galaxy S 3 = 1284

I know from first hand experience that the Galaxy S3 outperforms the Nexus in most intensive games and applicatkons so I won't put too much credence in the Geekbench score. The results of compared to the iPhone 5 have stirred up quite some discussion and chest thumping tho.

Good times at MacRumors. :D

matttye
Sep 16, 2012, 04:28 PM
As any person that has used android for awhile can tell you the OS requires heaps of high powered silicon to run well. It's a shoddy mobile OS.

It's not efficient and it doesn't scale worth a damn, so no, a jellybean ROM wont solve your problems.

Either way, I don't like Geekbench. I prefer seeing Anandtech's in depth look, but in the end, it really doesn't matter.

I'm never buying another android device, and my iPhone 5 should be here on Friday.

Jelly Bean is a step in the right direction. It's just as smooth as my iPad 2.

Looking forward to iOS 6 though, do they usually release it alongside the new iPhone?

ChazUK
Sep 16, 2012, 04:33 PM
It's not efficient and it doesn't scale worth a damn, so no, a jellybean ROM wont solve your problems.

.

Any reason you say this? My first hand experience with Jellybean on the Galaxy Nexus and 2 Tegra3 based tablets (Transformer Prime on ICS and Nexus 7 on JB) have seen massive performance improvements on the same silicone and architecture.

Given the weaker Tegra3 specification of the Nexus 7 over the Transformer, I'm surprised further by what Google have managed with JB.

Congratulations on ordering your new phone.

matttye
Sep 16, 2012, 04:38 PM
Any reason you say this? My first hand experience with Jellybean on the Galaxy Nexus and 2 Tegra3 based tablets (Transformer Prime on ICS and Nexus 7 on JB) have seen massive performance improvements on the same silicone and architecture.

Given the weaker Tegra3 specification of the Nexus 7 over the Transformer, I'm surprised further by what Google have managed with JB.

Congratulations on ordering your new phone.

I'm using a leaked version of Jelly Bean on my S3 and it's a lot smoother all round. Biggest performance gains can be seen in the browser.

Tapatalk is still a little bit laggy when it scrolls. Seems to be a badly coded app!

Technarchy
Sep 16, 2012, 04:50 PM
Jelly Bean is a step in the right direction. It's just as smooth as my iPad 2.

Looking forward to iOS 6 though, do they usually release it alongside the new iPhone?

JB is a step in the right direction, the problem is it isn't portable and doesn't scale very well.

Sure, it's fine on a GS3 international, but only 1% of android phones feature that caliber of hardware. The cost of entry is still high because android at its core is resource hungry, so when people say JB fixes everything, what they really should be saying is JB helps if you already have an expensive android phone.

Edit: For what its worth, if it isn't Exynos, I don't consider it a real GSII or GSIII.

matttye
Sep 16, 2012, 04:59 PM
JB is a step in the right direction, the problem is it isn't portable and doesn't scale very well.

Sure, it's fine on a GS3 international, but only 1% of android phones feature that caliber of hardware. The cost of entry is still high because android at its core is resource hungry, so when people say JB fixes everything, what they really should be saying is JB helps if you already have a well above average android phone.

Edit: For what its worth, if it isn't Exynos, I don't consider it a real GSII or GSIII.

The Galaxy Nexus has a dual core processor clocked at 1.2GHz which was a little slower than the Exynos in the Galaxy S2 based on benchmarks, so it seems the last generation of flagship Androids could run Jelly Bean.

What exactly is so resource hungry about Jelly Bean that it will only run on the best hardware? Given the performance gains the Nexus has seen by running it, I'd say it's evident that it works on older hardware. Unless you'd say that the Nexus is in the top 1% of hardware?

How about the HTC Desire (http://forum.xda-developers.com/showthread.php?t=1827934)? :p That's quite ancient now and appears to be quite capable of running Jelly Bean.

Chrisg2014
Sep 16, 2012, 07:59 PM
It's the software. Like it has been said so many times. But also I think that apple doesn't advertise this is because if you go up to someone and say. Hey do you want a computer that has (making this up here) 3GB of ram, 250GB of storage and a 1.5ghz processer or one that is double the specs and at the same price. Most people would go with the one that has more to offer. If you don't tell them whats under the hood, people won't know that you actually over payed for a computer.

Vegastouch
Sep 16, 2012, 08:03 PM
Yes, but this seriously results in a degraded battery life.

Actually it doesnt. The GS3 has a better battery life than the 4S while powering a screen 1.3" bigger.

And get your specs straight. The UK GS3 has a 1.4 Ghz quadcore with 1 GB of RAM.

US version has a 1.5 Ghz Dual core, 2GB of RAM.

TheMacBookPro
Sep 16, 2012, 08:06 PM
I'm never buying another android device, and my iPhone 5 should be here on Friday.

Ooooh, congratulations. We're all very excited for you and I'm sure that Samsung/HTC/etc are going to miss the sales that you provided for them in the past very very much :rolleyes:

irDigital0l
Sep 16, 2012, 08:08 PM
S3 with Jellybean still better than iphone 5 though according to Geekbench...

Technarchy
Sep 16, 2012, 08:17 PM
Ooooh, congratulations. We're all very excited for you and I'm sure that Samsung/HTC/etc are going to miss the sales that you provided for them in the past very very much :rolleyes:

And I'm sure Apple appreciates the sale, so there you go.

gwelmarten
Sep 17, 2012, 03:55 AM
Actually it doesnt. The GS3 has a better battery life than the 4S while powering a screen 1.3" bigger.

And get your specs straight. The UK GS3 has a 1.4 Ghz quadcore with 1 GB of RAM.

US version has a 1.5 Ghz Dual core, 2GB of RAM.

I was referring to the specs published on the GeekBench page for the phone. But you are right. So instead of it being 4+ x the power, it's 2.5+ x the power. Which is still a lot they are wasting. Idiots.

----------

Really? This is what I get when I search:

Iphone 5 (1601 points):
http://browser.primatelabs.com/geekbench2/1030202

Galaxy S3 4-core, 1GB (Non-US version) (1847 points):
http://browser.primatelabs.com/geekbench2/998950

Galaxy S3 2-core, 2GB (US version) (1683 points):
http://browser.primatelabs.com/geekbench2/884649

That said, pure processing power isn't necessarily a good measuring stick for how a cell phone performs in day-to-day activities.

Click the link in the MacRumors home page story about the geek bench score for the iPhone 5 - it says 1560 as an average for all S3 devices on the all Android devices page.

May I ask, why do Samsung see it necessary to make two versions of the phone anyway? Do they see the UK market as having different needs to the US one, or is this just some blown up example of A/B testing?

nba1341
Sep 17, 2012, 04:54 AM
I was referring to the specs published on the GeekBench page for the phone. But you are right. So instead of it being 4+ x the power, it's 2.5+ x the power. Which is still a lot they are wasting. Idiots.

----------



Click the link in the MacRumors home page story about the geek bench score for the iPhone 5 - it says 1560 as an average for all S3 devices on the all Android devices page.

May I ask, why do Samsung see it necessary to make two versions of the phone anyway? Do they see the UK market as having different needs to the US one, or is this just some blown up example of A/B testing?

U.S. networks did not support quad-core at the time. Same thing happened with One X

KnightWRX
Sep 17, 2012, 06:33 AM
Guys, Quad Core Cortex A9 vs Dual Core Cortex A15.

There's nothing magical about OS/hardware integration here, this is simply about Apple having used the latest ARMv7 architecture that can run more instructions per clock than the older architecture.

That's it. No magic here. This is Core 2 Duo vs Sandy Bridge, ARM style.

JB is a step in the right direction, the problem is it isn't portable and doesn't scale very well.

You don't even know the definitions of the words you're using right ? Portable ? Scale ?

What's not portable about Android ? It runs a VM on top of one of the most portable OS kernels around, Linux, and as such can run on multiple CPU architectures without requiring all the applications in the Google Play store to be recompiled to the new architecture. There are already x86 phones (Xolo X900) out there based on Intel's SoC (Medfield). The only thing that requires porting is Davlik itself.

As for scaling ? What are you on about ? The fact that the Exynos 4412 Quad in the SGSIII scores about double what the iPhone 4S, both based on Cortex A9 architectures should tell you all you need to know about scaling : the Linux kernel is one of the best scaling operating system out there. That's why it can run your phone, your TV, your watch and at the same time it can run multi-terabyte SAN switching equipement, mainframe type computers and superdome style clustering environnements.

No seriously, stick to what you know if you want to bash Android, don't bash areas of it that you obviously have no clue about, especially using words that you don't know the definition of.

Anyway, I thought you didn't participate in Android threads since you don't like it ? Why are you wasting time reading and responding to Android threads if the OS doesn't interest you ? I don't go around Windows Phone forums to post if I'm not interested in the product... Are you one of those guys ?

----------

It's the software. Like it has been said so many times.

And it has been wrong all the times its been said. The Darwin operating system is not anymore efficient at dispatching the type of code that Geekbench runs than Linux is.

You guys need to understand what Geekbench actually is : it's a synthetic benchmark designed to run as many instructions as possible on as many cores as possible without having to wait on any other code. It's just designed to saturate the processor as much as possible.

The OS doesn't really factor in here. This is purely about Cortex A9 vs Cortex A15 and instructions per clock. Anyone who doesn't realise this and thinks it's Apple pixie dust sure doesn't understand a whole out about how computers work.

baltoyyz
Sep 17, 2012, 06:34 AM
I was referring to the specs published on the GeekBench page for the phone. But you are right. So instead of it being 4+ x the power, it's 2.5+ x the power. Which is still a lot they are wasting. Idiots.

----------



May I ask, why do Samsung see it necessary to make two versions of the phone anyway? Do they see the UK market as having different needs to the US one, or is this just some blown up example of A/B testing?

The battery drain with the quad core and lte would have been to great. So they released the quad core without lte and then dual core with lte. They also added an extra gig of ram in the dual core to make up some speed difference.

Tarzanman
Sep 17, 2012, 07:26 AM
What are Samsung doing with all this extra power?

DLNA
USB host mode
Android beam / NFC
Bluetooth 4.0
Native DivX/Xvid/h.264 AVI/MKV support
Wolfson DACs

et cetera.

zbarvian
Sep 17, 2012, 07:40 AM
Ok, this should be in the main thread, but what the hell. I'll bite. Android isn't a resource friendly OS. Much like Windows manufacturers advertise heavily on specs where as Apple doesn't need that much. Reason?

Android is a general OS for all types of phones. So it is in a sense bloated and has to account for several types of configurations as possible.

iPhone OS, is custom tailored to just one set of hardware. Hell, iPhone OS 6.0 dropped the support for iPhone and iPhone 3G in its entirety. Only support for the iPhone 3GS, iPhone 4 and iPhone 4S remain. Apple continuously optimizes iPhone OS to make it as fast and performance efficient with as little as possible.

Result? S3 is beaten in Geekbench scores.

This guy has it right.

KnightWRX
Sep 17, 2012, 07:59 AM
This guy has it right.

Nope, he doesn't. He's just repeating the same spiel that everyone is. It's about CPU architectures, not about OS differences.

Zaft
Sep 17, 2012, 08:29 AM
Who cares about the benchmarks. The Galaxy S3 is the smoothest android device iv encountered. Even smoother then a Galaxy Nexus running JB from my experience.

daveathall
Sep 17, 2012, 08:45 AM
I never knew about these benchmark tests before yesterday. Must admit, it hasn't diminished my enjoyment when using my SGS3, it is as smooth and as fast as my 4S, not that I have or am capable of timing it, it just seems that way, which is good enough for me.

I do wonder about these tests though. As a casual user that doesn't multi task that much, does it really matter that this or that phone is faster by a factor of 41 than the latest offerings? Would I even notice the difference? I don't think so and wouldn't lose any sleep if my phone was a millisecond slower than another.

zbarvian
Sep 17, 2012, 08:57 AM
Nope, he doesn't. He's just repeating the same spiel that everyone is. It's about CPU architectures, not about OS differences.

So is that why a 1.5 GHz Krait (better architecture) also performs worse?

jav6454
Sep 17, 2012, 09:04 AM
Nope, he doesn't. He's just repeating the same spiel that everyone is. It's about CPU architectures, not about OS differences.

Not withstanding I stated that long before the 5 people before me posted. Given that it is always the case when writing long. At any rate, Ed part of my response is what has been known for a long time.

KnightWRX
Sep 17, 2012, 11:16 AM
So is that why a 1.5 GHz Krait (better architecture) also performs worse?

Krait is not as good as Cortex A15 :

http://en.wikipedia.org/wiki/Krait_%28CPU%29

Again folks, instructions per clock. Mhz and cores aren't the only thing that changes performance of a chip.

Technarchy
Sep 17, 2012, 11:57 AM
As for scaling ? What are you on about ? The fact that the Exynos 4412 Quad in the SGSIII scores about double what the iPhone 4S, both based on Cortex A9 architectures should tell you all you need to know about scaling : the Linux kernel is one of the best scaling operating system out there. That's why it can run your phone, your TV, your watch and at the same time it can run multi-terabyte SAN switching equipement, mainframe type computers and superdome style clustering environnements.



It scales so well that android does not have a reputation for laggy performance...

zbarvian
Sep 17, 2012, 12:14 PM
Krait is not as good as Cortex A15 :

http://en.wikipedia.org/wiki/Krait_%28CPU%29

Again folks, instructions per clock. Mhz and cores aren't the only thing that changes performance of a chip.

And the A6 is? If the A6 at 1 GHz is substantially faster than a 1.5 GHz Krait, the architecture of the A6 must be far more advanced than an A15 (since A15 is only 40% faster than A9). Not to mention, Krait is a hybrid between an A9 and A15, so it has better architecture than an A9 already. By all accounts, a 1.5 GHz Krait should be significantly quicker than a 1 GHz A6. Face it, Android is not as well optimized for the chipset, and it yields worse performance. There's really no excuse.

0m3ga
Sep 17, 2012, 12:25 PM
It scales so well that android does not have a reputation for laggy performance...

Android being made for all different sized screens and hardware variations is inherently better at scaling apps than iOS. Developers have multiple options to choose from, in terms of how they want their apps to look on various hardware. Apple has also had a similar option built in to their OS for 2 years now, but there has been no incentive for developers to use this option. So, until iOS developers completely rewrite their apps, or make new ones, this is wha5 you need to get accustomed to on the iphone 5. Android doesn't have this issue, because the OS was designed to compensate from the beginning.

Iphone 4S on top, and iphone 5's on the bottom. The white iphone really makes the letterboxing stand out.
http://cdn1.sbnation.com/imported_assets/1190617/29as8hy.jpg

Technarchy
Sep 17, 2012, 12:29 PM
Android being made for all different sized screens and hardware variations is inherently better at scaling apps than iOS. Developers have multiple options to choose from, in terms of how they want their apps to look on various hardware. Apple has also had a similar option built in to their OS for 2 years now, but there has been no incentive for developers to use this option. So, until iOS developers completely rewrite their apps, or make new ones, this is wha5 you need to get accustomed to on the iphone 5. Android doesn't have this issue, because the OS was designed to compensate from the beginning.

Iphone 4S on top, and iphone 5's on the bottom. The white iphone really makes the letterboxing stand out.
Image (http://cdn1.sbnation.com/imported_assets/1190617/29as8hy.jpg)

So?

0m3ga
Sep 17, 2012, 12:38 PM
So?

Weren't you saying that Android doesn't scale well? It does in today's phones running 4.0 and higher. My S3 scales everything beautifully, whether the app is old and designed for a 4" screen or an app designed for 5he Kindle Fire and coming from Amazons app store.

The iphone doesn't hav3 a scaling problem, BECAUSE IT DOESN'T SCALE AT ALL. It is awesome that you have a 4 inch screen. To bad none of your apps will be able to use the extra real estate. Might as well just stayed with the 4S. :D

Carouser
Sep 17, 2012, 12:39 PM
So?

The argument amounts to "Android developers have been working with scaling apps, because that's what the platform demands. iOS developers have been working with fixed dimensions, because that's what the platform demands. Therefore, iOS apps don't scale well! Android wins this round, heh."

It ignores the fact that scaling is a dumb solution to the new iPhone/iPod resolution, and isn't a virtue in and of itself. Now iOS developers have to accommodate one more resolution for a few million more potential customers. Big deal.

EDIT: And what ChazUK pointed out below, of course.

ChazUK
Sep 17, 2012, 12:46 PM
Weren't you saying that Android doesn't scale well? It does in today's phones running 4.0 and higher. My S3 scales everything beautifully, whether the app is old and designed for a 4" screen or an app designed for 5he Kindle Fire and coming from Amazons app store.

The iphone doesn't hav3 a scaling problem, BECAUSE IT DOESN'T SCALE AT ALL. It is awesome that you have a 4 inch screen. To bad none of your apps will be able to use the extra real estate. Might as well just stayed with the 4S. :D

I think he is referring to scaling on low end hardware like how Apple has managed to get iOS to perform well on phones like the 3GS/iPhone 4 but similar specced phones will struggle to run the latest version of Android.

Technarchy
Sep 17, 2012, 12:56 PM
I think he is referring to scaling on low end hardware like how Apple has managed to get iOS to perform well on phones like the 3GS/iPhone 4 but similar specced phones will struggle to run the latest version of Android.

You got it. Cheerio.

matttye
Sep 17, 2012, 02:57 PM
You got it. Cheerio.

You've completely ignored my post which pointed out that the HTC desire can run jelly bean.

gwelmarten
Sep 17, 2012, 03:47 PM
Hi Everyone

I noticed whilst on this thread that their seemed to be lots of Samsung supporters here. Something I didn't expect.

I'd be interested to hear where you stand on the Apple VS Samsung issue. Vote here (http://forums.macrumors.com/showthread.php?t=1444258).

Sam

daveathall
Sep 17, 2012, 03:55 PM
Hi Everyone

I noticed whilst on this thread that their seemed to be lots of Samsung supporters here. Something I didn't expect.

I'd be interested to hear where you stand on the Apple VS Samsung issue. Vote here (http://forums.macrumors.com/showthread.php?t=1444258).

Sam

You are obviously a big fan of Apple, nothing wrong with that, but why do you keep posting in this forum?

Which is surprising to me - as this is MacRumors - a site where you'd expect the majority of people to be Apple fans (or at least interested in Apple).


Have you noticed this;

http://i219.photobucket.com/albums/cc236/daveathall/ScreenShot2012-09-17at182352.png

Alternative to iOS and iOS devices. Do you know what that means?

Also, have a look at my sig line.

pear21
Sep 17, 2012, 04:52 PM
I thought the us version of the s3 was dual core and the international was quad?

Yes that is correct so that means the s3 that people use in the US aren't much faster than an iPhone in my opinion and that's why I'm excited to test out an even faster iphone

Mrg02d
Sep 17, 2012, 05:41 PM
Yes that is correct so that means the s3 that people use in the US aren't much faster than an iPhone in my opinion and that's why I'm excited to test out an even faster iphone

Except it ISN'T faster, is it? Plenty of tests came out to show that the s3 is faster than the 5. 1800 or whatever at 1500Mhz is still faster than 1601, isn't it? You can't clock the 5 higher, so there you go bro.

pear21
Sep 17, 2012, 07:07 PM
Except it ISN'T faster, is it? Plenty of tests came out to show that the s3 is faster than the 5. 1800 or whatever at 1500Mhz is still faster than 1601, isn't it? You can't clock the 5 higher, so there you go bro.

Will you ever notice that small of a difference though? No but you will say oh yeah I totally can notice bro.

Mrg02d
Sep 17, 2012, 09:02 PM
Will you ever notice that small of a difference though? No but you will say oh yeah I totally can notice bro.

I can feel when apple fanatics need to be sat down, thats for sure. Getting all ape like over erroneous results is ridiculous. Pages and pages of garbage the other night.

Its like watching a touchdown get called back. :D

pear21
Sep 17, 2012, 09:19 PM
I can feel when apple fanatics need to be sat down, thats for sure. Getting all ape like over erroneous results is ridiculous. Pages and pages of garbage the other night.

Its like watching a touchdown get called back. :D

That doesn't even make sense

Mrg02d
Sep 17, 2012, 09:23 PM
That doesn't even make sense

Exactly. Glad we came together on this, amiga.

watchthisspace
Sep 17, 2012, 09:24 PM
I'd like to point out, that as far as I'm aware of, the Geekbench results don't include the set of up of the devices. The iPhone for all we know, could be a fresh restore with only the Geekbench app, while the s3, could be running lots of apps in the background, or be stripped down.

That's the only thing I really don't like about these results.

Wait until the iPhone 5 is, and hopefully someone will have a video showing some objective benchmarks of the iPhone and S3 doing something. Even then the results are quite Apple's to Oranges because both devices can't run the exact same set up.

Mrg02d
Sep 17, 2012, 09:29 PM
I'd like to point out, that as far as I'm aware of, the Geekbench results don't include the set of up of the devices. The iPhone for all we know, could be a fresh restore with only the Geekbench app, while the s3, could be running lots of apps in the background, or be stripped down.

That's the only thing I really don't like about these results.

Wait until the iPhone 5 is, and hopefully someone will have a video showing some objective benchmarks of the iPhone and S3 doing something. Even then the results are quite Apple's to Oranges because both devices can't run the exact same set up.

Yup. That would explain why the 5 was bested by so many s3 users shortly after that crap was put out there.

lazard
Sep 17, 2012, 09:51 PM
Except it ISN'T faster, is it? Plenty of tests came out to show that the s3 is faster than the 5. 1800 or whatever at 1500Mhz is still faster than 1601, isn't it? You can't clock the 5 higher, so there you go bro.

http://browser.primatelabs.com/geekbench2/1038481

Mrg02d
Sep 17, 2012, 09:55 PM
http://browser.primatelabs.com/geekbench2/1038481

Whoa! So much for the 5 being all revolutionary.:rolleyes:

I'm guessing you are running cm10? Which nightly you on?

KentuckyHouse
Sep 17, 2012, 09:57 PM
Yes, but this seriously results in a degraded battery life.

No, it doesn't. Again, bring facts or don't comment on what you don't know.

KentuckyHouse
Sep 17, 2012, 09:59 PM
I thought the us version of the s3 was dual core and the international was quad?

The US version is dual core with 2gb of RAM. International version is quad core with 1gb of RAM.

Mrg02d
Sep 17, 2012, 10:03 PM
What's with all this talk of degraded battery life? My s3 lasts longer than ANY of my other smartphones...Sounds like regurgitated barf to me.

roxxette
Sep 18, 2012, 05:02 AM
i checked some results on geekbench, i dont know why its advertised the s3 score of 15xx when i see a lot of scores of 17xx, 18xx and higher :confused:

dontpannic
Sep 18, 2012, 06:39 AM
Except it ISN'T faster, is it? Plenty of tests came out to show that the s3 is faster than the 5. 1800 or whatever at 1500Mhz is still faster than 1601, isn't it? You can't clock the 5 higher, so there you go bro.

Overclocking is cheating.

KnightWRX
Sep 18, 2012, 07:28 AM
It scales so well that android does not have a reputation for laggy performance...

Laggy performance (UI lag) is not an indication of lack of scaling or scaling. You do not understand the word if that is what you think, I'm sorry to say.

UI lag can come from many sources, but the main source is processing UI redrawing on the main thread, thus blocking all other events until you're done rendering your final buffer. Usually, this is where you will want to leverage your thread scheduler by offloading the rendering of the back buffer to a different thread, leaving your main thread free to still process events and prepare the datasets for the next UI frame. When the backbuffer is done rendering, you simply use the main thread to flip the buffers.

You have to realise, to run anything at 60 fps, you have to have your rendered buffers ready in 16 ms each time.

All of this, as you probably don't know, has nothing whatsoever to do with scaling. Nothing. Nada.

Scaling is the function of simply being able to scale to bigger or lesser configuration. If you have a performance of 100 on a certain processor by a certain metric, linear scaling will mean you have 200 on twice that processor and 50 on half of it. Linear scaling is most impossible to reach due to overhead, but the Darwin and Linux kernels are both very good at scaling and reducing that overhead, on both sides. That means that the OS is capable of utilizing the extra hardware you throw at it very well as to make your investment worthwhile and will gracefully give you what you expect of lesser hardware.

But really, I don't think you really are here to learn or actually use the proper language. You just want to bash Android don't you ?

----------

I think he is referring to scaling on low end hardware like how Apple has managed to get iOS to perform well on phones like the 3GS/iPhone 4 but similar specced phones will struggle to run the latest version of Android.

Which has nothing to do with the "OS scaling" at all. That's not what scaling means or is. Anyway, you and I both know what Techanarchy is here to do, I don't know why we bother with him anymore. The guy is basically doing what he accused others of doing in the iPhone forums.

Goes to show where the real troublemakers were. Guy is so obsessed with his vendetta against Android, he can't stop himself from posting in threads about it. Of course, anyone else that doesn't like something wouldn't bother wasting all their energy on it...

lazard
Sep 18, 2012, 07:57 AM
Overclocking is cheating.

2087 @ stock 1.4GHz

http://browser.primatelabs.com/geekbench2/1038481

swy05
Sep 18, 2012, 08:54 AM
The US version is dual core with 2gb of RAM. International version is quad core with 1gb of RAM.

Or the quad core 2gb korean version.

This phone is fantastic =P

dontpannic
Sep 18, 2012, 09:30 AM
2087 @ stock 1.4GHz

http://browser.primatelabs.com/geekbench2/1038481

I was specifically quoting the guy I quoted due to his comment "you can clock it higher".

Mrg02d
Sep 18, 2012, 09:34 AM
Overclocking is cheating.

No one is overclocking for those results Brah. Face it, your iPhone 5 doesn't beat the s3. Shows over...

cotak
Sep 18, 2012, 09:57 AM
Laggy performance (UI lag) is not an indication of lack of scaling or scaling. You do not understand the word if that is what you think, I'm sorry to say.

UI lag can come from many sources, but the main source is processing UI redrawing on the main thread, thus blocking all other events until you're done rendering your final buffer. Usually, this is where you will want to leverage your thread scheduler by offloading the rendering of the back buffer to a different thread, leaving your main thread free to still process events and prepare the datasets for the next UI frame. When the backbuffer is done rendering, you simply use the main thread to flip the buffers.

You have to realise, to run anything at 60 fps, you have to have your rendered buffers ready in 16 ms each time.

All of this, as you probably don't know, has nothing whatsoever to do with scaling. Nothing. Nada.

Scaling is the function of simply being able to scale to bigger or lesser configuration. If you have a performance of 100 on a certain processor by a certain metric, linear scaling will mean you have 200 on twice that processor and 50 on half of it. Linear scaling is most impossible to reach due to overhead, but the Darwin and Linux kernels are both very good at scaling and reducing that overhead, on both sides. That means that the OS is capable of utilizing the extra hardware you throw at it very well as to make your investment worthwhile and will gracefully give you what you expect of lesser hardware.

But really, I don't think you really are here to learn or actually use the proper language. You just want to bash Android don't you ?

----------



Which has nothing to do with the "OS scaling" at all. That's not what scaling means or is. Anyway, you and I both know what Techanarchy is here to do, I don't know why we bother with him anymore. The guy is basically doing what he accused others of doing in the iPhone forums.

Goes to show where the real troublemakers were. Guy is so obsessed with his vendetta against Android, he can't stop himself from posting in threads about it. Of course, anyone else that doesn't like something wouldn't bother wasting all their energy on it...

The whole scaling thing about iOS is stupid. Really stupid because if you actually look at what they do, they give you less and less of the new features so it'll run ok on the older hardware. So yes you might get a bigger numerical version but you aren't getting all the new features. Sometimes the new features are not given even when the hardware should be able to run it, all to drive sales. After all Apple makes almost all it's money from hardware sales. If you need to buy a new phone to get all the new features, it's no different from Android where some manufacturers aren't good at providing updates. In either camp you can't sit on a phone for decades and expect to have the latest and greatest software.

As for Android performance. VM code is not really a major performance bottleneck. If VM code is that bad Java would not be so popular. Yes there's an overhead but it's a lot less than most people think. And the ability to do some optimization with VM code which you cannot with native code (such as C++) means that in some situation it's actually faster (http://blog.cfelde.com/2010/06/c-vs-java-performance/). Before you boys shoot your mouth of try asking Siri these question eh? (Or just google, it might be quicker typing).

The whole Java VM vs native code thing shows just how little some guys who wrote in this thread knows about the devices they are using. Much as you like to think you are very clever most of you haven't faced and pass the sort of interviews the leading tech firms have for hiring their employees. Believe me, the guys who works for Apple and Google knows their stuff. They don't make silly choices like using a VM when it's clearly a lot slower.

So why is Android seemingly slower? Well, first of all they decided to do a iterative process vs what apple did with the iPhone. Being late to market means they either do what RIM is choosing to do (and losing users because of it) and waiting to release a "perfect" product. Or they can release something that sort of works and iterate on it to get better. Considering their market share they made the right choice. And the latest Android version is perfectly fast and usable.

As for the benches, the S3 comes in faster in some test (using the same bench) and then in other tests slower. Which makes sense considering that Android has real multi-tasking. So I wouldn't make a judgement till some more reliable testing gets done.

cynics
Sep 18, 2012, 10:31 AM
Only thing I gathered from this thread is Apple guys are really into specs. A lot of these types of threads are being started by people that prefer Apple.

My outlook on it is end results. My 4S is smoother then some Android devices. But at the same time the Android device is doing more (widgets, live wallpapers, certain apps running in the back ground, etc). It's just user preference in the end, smooth scrolling vs more content aka form over function. Now with the latest version of Android it's nearly or as smooth with still having all the function.

Point being, I don't rely on benchmarks, they don't mean crap. I rely on how it works for me.

dontpannic
Sep 18, 2012, 10:43 AM
No one is overclocking for those results Brah. Face it, your iPhone 5 doesn't beat the s3. Shows over...

Who's talking about results? I never said the S3 was worse than the iPhone 5 "brah". All I said was 'overclocking is cheating' to the guy who said 'you can clock the s3 higher'. Nothing more, "brah".

NZed
Sep 18, 2012, 11:35 AM
Full multitasking, widgets, smart stay, numerous gestures, application voice commands,... These are just a few of the things constantly running on the S3 that aren't on the iPhone.

This

and Android too

Evoken
Sep 19, 2012, 06:37 AM
The Geekbench results showed that the Galaxy S3 running the latest version of Android is a lot faster than the iPhone 5 running the latest version of iOS (1780 for the S3 vs 1600 for the iPhone). A better optimized OS allows the power of the S3 to really shine. Can't wait to get Jelly Bean next month! It will turn my S3 into quite a beast :D