Nintendo
Nintendo Switch
Last edited by a moderator:
Apple iPhone 4S dominates other smartphones
- Thread starter Nintendo
- Start date
If anyone from the UK is buying a 4S, and wants to go with Three, give me a shout for my affiliate link! I can probably hook you up with some casheroonie back. 
All the folks who said the iPhone 4S was fail have a lot of hat to eat. People were far to quick to judge and condemn it for something so superficial as not having an aesthetic redesign. Pah!
I, for one, *LIKE* the look of the iPhone 4/4S and didn't want it to change. A larger edge-to-edge screen wouldn't have gone amis though.
All the folks who said the iPhone 4S was fail have a lot of hat to eat. People were far to quick to judge and condemn it for something so superficial as not having an aesthetic redesign. Pah!
I, for one, *LIKE* the look of the iPhone 4/4S and didn't want it to change. A larger edge-to-edge screen wouldn't have gone amis though.
craigix
Mega GP Mania
It may well be really fast, but it's still running iOS which is well behind Android at the moment.
Real story, from Anandtech, as expected the A5 is clocked down so the A9 dual core is running at 800MHz with everything else likely throttled down respectively. The large die they used associated with their use of a PowerVR SGX 543MP2 is going to intrinsically be a serious power concern when you actually have significant portions of the die powered on. Something that won't be helped by it using a LG LCD screen when its "nearest competitor" uses a Super AMOLED Plus screen.
Android 4.0 aka "Ice Cream Sandwich" is only currently delayed in the sense it's launch event has been held off because Google and associated have enough honor to respect the death of the long time high profile CEO of one of their main competitors.
None of these tests were actually conducted by Anandtech, although the results aren't unexpected. Given the games that I hear about involve a 2D compositor instead of a 3D engine, Texas Instruments efficiency add from having a dedicated 2D compositor shouldn't be ignored.
Android 4.0 aka "Ice Cream Sandwich" is only currently delayed in the sense it's launch event has been held off because Google and associated have enough honor to respect the death of the long time high profile CEO of one of their main competitors.
None of these tests were actually conducted by Anandtech, although the results aren't unexpected. Given the games that I hear about involve a 2D compositor instead of a 3D engine, Texas Instruments efficiency add from having a dedicated 2D compositor shouldn't be ignored.
Last edited by a moderator:
> It may well be really fast, but it's still running iOS which is well behind Android at the moment.
I've been using a Galaxy S since my iPhone 4 went down a few weeks ago - iOS is cooler / smoother / slicker than Android. But it's not open of course if not jail-broken. But for what I want to do on a handheld phone ( call, txt, a bit of surfing ), I much prefer iOS - both sit on top of unix, but with Android you definitelt notice it - with iOS you totally don't.
Of course, imho
Mal
I've been using a Galaxy S since my iPhone 4 went down a few weeks ago - iOS is cooler / smoother / slicker than Android. But it's not open of course if not jail-broken. But for what I want to do on a handheld phone ( call, txt, a bit of surfing ), I much prefer iOS - both sit on top of unix, but with Android you definitelt notice it - with iOS you totally don't.
Of course, imho
Mal
> It may well be really fast, but it's still running iOS which is well behind Android at the moment.
I've been using a Galaxy S since my iPhone 4 went down a few weeks ago - iOS is cooler / smoother / slicker than Android. But it's not open of course if not jail-broken. But for what I want to do on a handheld phone ( call, txt, a bit of surfing ), I much prefer iOS - both sit on top of unix, but with Android you definitelt notice it - with iOS you totally don't.
Of course, imho
Mal
So in other words you consider Unix OS guts uncool and want to be shielded from them... On a forum related to a Linux device. I'll give you points for audacity, but there's a reason there's this term n00b that's differentiated from newb and why one has a lot of scorn related to it.
sporkbox
Member
- Joined
- Jan 10, 2011
- Messages
- 186
- Age
- 38
Still not buying one. I've used both iOS and Android and neither are all that great. I'll stick to plain-old Linux without mechanisms that try to limit what I can do.
pre-empt: If I have to break the limitations, it's designed against me to begin with.
pre-empt: If I have to break the limitations, it's designed against me to begin with.
Last edited by a moderator:
Phelsumania
Member
well how has the htc hd2 not made this list?? mine overclocks to 1.3ghz and is fast as hell. and surely the new htc titan with a 1.5ghz processor was to be overclocked it would smash a gay iphone out the water. i would even go as far to say am sure an overclocked hd2 which is quite a few years old now would batter the new Gayphone (having never bothered checking the specs out of the new iphone)
Phelsumania
Member
looking back at this actually surely if i am correct in thinking the titan does have a 1.5ghz processor (am sure ive read on the htc website it does) then its going to batter the iphone without any overclocking involved anyway.
Jarvis:
Here's the thing. Apple with their A5 processor decided to put more of an emphasis on the GPU side of things then other ARM chip vendors did.
As a result they're currently the one of the few to go far enough up PowerVR's product line to use the SGX 543/544 at all, and the only one putting it in a multicore configuration with this process. Their use of a SGX 543MP2 as a result represents the most potent mobile GPUs currently available. Based on the wiki numbers their GPU is theoretically able to put out, at 200MHz, 70 Million Triangles/s, 2GPixel/s, and 14.4 GFLOP/S.
So while on the pure CPU side of things the iPhone 4S only has an advantage over a Cortex A8 running at 1GHz when it can exploit multicore or out of order advantages, it does have a relatively powerful GPU backing it up. An OMAP 4430 or 4460 for instance is running a single SGX 540 for its GPU section resulting in it having a theoretical 28 Million Triangles/s, 1GPixel/s, and 3.2 GFLOPS at 200 MHz, ignoring NEON contributions. The flipside of this is that the GPU power difference is only really important with OpenCL accelerated applications and 3D games. If you're playing 2D games, like Angry Birds, the Apple A5 is going to be consuming more power then an OMAP 4 processor however because it is going to need to power more of the die up and which will result in more loses. The LG "Retina" LCD display used is also going to tend to consume more power then the increasingly refined AMOLED technology present in Android handsets.
Basically most of the current handset GPUs are in the territory of 6th generation consoles, while Apple's A5 should cleanly dominate them in the GPU department at least. Nothing will be on the level of theoretical throughput/power with 7th generation consoles until refined 28nm parts come out based on what's currently known.
Here's the thing. Apple with their A5 processor decided to put more of an emphasis on the GPU side of things then other ARM chip vendors did.
As a result they're currently the one of the few to go far enough up PowerVR's product line to use the SGX 543/544 at all, and the only one putting it in a multicore configuration with this process. Their use of a SGX 543MP2 as a result represents the most potent mobile GPUs currently available. Based on the wiki numbers their GPU is theoretically able to put out, at 200MHz, 70 Million Triangles/s, 2GPixel/s, and 14.4 GFLOP/S.
So while on the pure CPU side of things the iPhone 4S only has an advantage over a Cortex A8 running at 1GHz when it can exploit multicore or out of order advantages, it does have a relatively powerful GPU backing it up. An OMAP 4430 or 4460 for instance is running a single SGX 540 for its GPU section resulting in it having a theoretical 28 Million Triangles/s, 1GPixel/s, and 3.2 GFLOPS at 200 MHz, ignoring NEON contributions. The flipside of this is that the GPU power difference is only really important with OpenCL accelerated applications and 3D games. If you're playing 2D games, like Angry Birds, the Apple A5 is going to be consuming more power then an OMAP 4 processor however because it is going to need to power more of the die up and which will result in more loses. The LG "Retina" LCD display used is also going to tend to consume more power then the increasingly refined AMOLED technology present in Android handsets.
Basically most of the current handset GPUs are in the territory of 6th generation consoles, while Apple's A5 should cleanly dominate them in the GPU department at least. Nothing will be on the level of theoretical throughput/power with 7th generation consoles until refined 28nm parts come out based on what's currently known.
Last edited by a moderator:
? You're not making much sense. ARM CPUs can idle, and the OMAP 4 line is designed to use M-series ARM processors and a 2D engine to run the desktop allowing the CPU and GPU sections to be gated off. The Apple A5 doesn't appear to have anything vaguely similar. Plus unless you're going heavy on the white AMOLEDs tend to be more efficient then basic LCDs like Apple is using.
If you mean the iPhone line while they apparently went from a 5.254 Whr to a 5.5 Whr battery battery, but the Apple A5 and A4 are both supposed to have been fabricated on Samsung's 45nm process. Hence Apple's specs indicate that idle time went from 300 hours with the IPhone 4 & 3GS to 200 hours for the iPhone 4S, with the iPhone 4S beating the iPhone 4 only in terms of 3G talk time 8 hours to 7.
Note the IPhone 4S' issues verse the IPhone 4 should be worse as the CPU and GPU die come off idle, and the DSP and other differable logic become less able to allow the rest of the die to be gated off. Well at least if you ignore race to execute efficiency adds, but those favor Android handsets more. I don't use either, but with the same usage profiles I don't really see an OMAP 4430 losing verse a A5 everything else being equal, which seems to pan out with the Droid Bionic which current reviews would indicate is superior to Apple's published battery life specs. Based on the reviews I'm seeing the iPhone 4S loses to the point the Droid Bionic's 4G talk time is effectively the same as the iPhone 4S' 3G talk time.
The iOS 5 update is only going two years back to the iPohne 3GS as far as I can tell, and that mainly seems to be because they're not shutting down that product line quite yet. Their iPad 1 vs 2 software situation on the other hand is a clear example that when they're ready to do a phase out that they don't care regardless of how long it's been.
So I'm going to have to say you're being overly optimistic, and seem to be trying to buck the actual trends. Android users seem to be the ones that just end up hanging onto their handsets.
Last edited by a moderator:
Exophase
Nothing good will ever come of Exophase.
Okay, so you're still harping on this theory that larger die size = higher intrinsic power draw even though in the real world a lot of design optimizations trade area for less power rather than more. And you're still pegging A5 as having a huge die even though the only data point for comparison you have is the Tegra 2; you still don't actually know the size of OMAP4430 or Exynos 4210. But then that choice for the big 543MP2 should at least let Apple turn significant portions of die off whenever it isn't running demanding games, since it can literally turn half of the GPU off.
There's also no data that suggests Super AMOLED is actually drawing less than current low power LCDs, that's all still largely hype.
What the lower clocks could be is a combination of Apple using a smaller battery than competitors (which makes sense given that they're all striving to be minimally thin and Apple is using a smaller screen than many) and Apple actually giving a shit about battery life.
Jebe said:
No it can't. The two M3s are there strictly as peripherals for helping in dedicated concurrent tasks, probably video encode/decode: this is actually nothing new for OMAP, yes even OMAP3 has an ARM9 no one talks about. The M3 is probably connected directly to nothing more than a tiny SRAM. It's an embedded/microcontroller profile CPU. It can't run your desktop apps. The instruction set isn't even compatible. It's Thumb-2 only and doesn't support most of the extensions since v4, at least not verbatim.
I seem to recall some marketing material about them being able to get away with using the M-series to gate off the A9s in times of low activity, but I'm not seeing it when I went looking to confirm so I'll concede that.
Yes, the real question is the number of active transistors, their clocking, and how long they stay at raised clocks as related to purposeful current draw and leakage. Plus other factors like voltage settings if we're really going to get into it. The thing is in case you didn't notice we're relying on teardowns which tend to not involve directly counting transistors or otherwise getting quite that far into the guts of the thing.
FYI the RIM Playbook has had teardowns, and those indicate that it's OMAP 4430 is sized with a 14x14mm package and a 8.2x8.5mm die. The Apple A5 in teardowns has been measured to have a 16.7x14.3mm package and a 12.1x10.1mm die.
Not that this changes the fact we've seen nothing indicating that Apple has a dedicated 2D compositor and hence needs to constantly have at least one of the SGX 543s constantly going at a low idle. There's also the fact this is all borne out by Apple's own specifications despite the battery bump they did going from the iPhone 4 to iPhone 4S in the idle characteristics. I'm not rampantly speculating so much as explaining what the official specifications show, which cannot be said for your appealing to ignorance and wildly hitting the "Objection!" button.
Yes, the real question is the number of active transistors, their clocking, and how long they stay at raised clocks as related to purposeful current draw and leakage. Plus other factors like voltage settings if we're really going to get into it. The thing is in case you didn't notice we're relying on teardowns which tend to not involve directly counting transistors or otherwise getting quite that far into the guts of the thing.
FYI the RIM Playbook has had teardowns, and those indicate that it's OMAP 4430 is sized with a 14x14mm package and a 8.2x8.5mm die. The Apple A5 in teardowns has been measured to have a 16.7x14.3mm package and a 12.1x10.1mm die.
Not that this changes the fact we've seen nothing indicating that Apple has a dedicated 2D compositor and hence needs to constantly have at least one of the SGX 543s constantly going at a low idle. There's also the fact this is all borne out by Apple's own specifications despite the battery bump they did going from the iPhone 4 to iPhone 4S in the idle characteristics. I'm not rampantly speculating so much as explaining what the official specifications show, which cannot be said for your appealing to ignorance and wildly hitting the "Objection!" button.
Last edited by a moderator:
Exophase
Nothing good will ever come of Exophase.
Hadn't seen there were xrays for OMAP4430. That does still leave out Exynos 4210, which I somehow don't think has the world's smallest die with a Mali-400MP4.
Of course A5 will have a higher peak power draw than A4 if it kept with two Cortex-A9 cores at 1GHz and the SGX543MP when they're made on the same process. That doesn't mean that other 2x Cortex-A9 SoCs in phones consume so much less, and no, a not even 5% increase in battery capacity doesn't do an awful lot to negate this. That said, you can't really jump to pinning a lower standby specification on the SoC instead of, for instance, the wireless hardware.
OMAP4430 and 4460 don't even have that 2D compositor you keep bringing up, that's a 4470 feature and I don't think I need to tell you that it isn't out yet. By the time we see it in products I wouldn't be especially surprised to see A6.
Of course A5 will have a higher peak power draw than A4 if it kept with two Cortex-A9 cores at 1GHz and the SGX543MP when they're made on the same process. That doesn't mean that other 2x Cortex-A9 SoCs in phones consume so much less, and no, a not even 5% increase in battery capacity doesn't do an awful lot to negate this. That said, you can't really jump to pinning a lower standby specification on the SoC instead of, for instance, the wireless hardware.
OMAP4430 and 4460 don't even have that 2D compositor you keep bringing up, that's a 4470 feature and I don't think I need to tell you that it isn't out yet. By the time we see it in products I wouldn't be especially surprised to see A6.