If A8 is dual core, good. Glad Apple isn't jumping on the "more cores for the hell of it" bandwagon. They know what theyre doing when it comes to chip design, which is why the A7 still blows everything else out of the water in most respects, even Samsung's 8 core mobile chips.
A performance enhanced A8 would be good enough for iPhone, however it won't do for iPad and some other things Apple could be working on. This is why I've stated before that I could see "X" variants again of Apples processors. With the X variants targeting higher performance machines, it would be easy for Apple to deliver a highly optimized cell phone chip.
Beyond that I've heard solid rumors that Apples goals this go around have been improved power efficiency. People might end up underwhelmed by the next iPhone processor if they don't value battery life.
Apple has starved people of reasonable hardware upgrades. They are so desperate for an upgrade to the hardware that even just adding additional ram is enough to satisfy them.
If A8 is dual core, good. Glad Apple isn't jumping on the "more cores for the hell of it" bandwagon. They know what theyre doing when it comes to chip design, which is why the A7 still blows everything else out of the water in most respects, even Samsung's 8 core mobile chips.
Whatever Apple do with the chip, they make the most out of it. Samsung's octa-core chip was nothing more than just a gimmick because it only worked 4 cores at the time. How pathetic! That's why Samsung decided to use SnapDragon quad-core chip in their GS4 handset for US market instead of their own Exynos Octa-core (only for European GS4).
Apple has starved people of reasonable hardware upgrades. They are so desperate for an upgrade to the hardware that even just adding additional ram is enough to satisfy them.
Man, it’s a good thing you don’t actually believe this; it’s one of the best pieces of trolling I’ve seen in a while.
I look at it this way it is all about power, power as in watts. Considering Apple has a ways to go yet just to keep A7 core feed there is much that can be improved performance wise with the dual core chip. This is why I see a dual core for the iPhone as a real possibility. They have to find a balance between thermal power and computational power.
That would be great for iPhone, maybe not so great for iPad, an enhanced AppleTV or other devices Apple has coming.
a mobile device is a i/o device. It's the GPUs that do most of the work. With a single user and strong constraints on power consumption, Dual core (one controlling the general interface, one controlling the logic of the the apps in compute state) is enough.
Enough for a cell phone yes, maybe not for an iPad or enhanced Apple TV.
As it is im not sure you have a strong understanding of Grand Central Dispatch and other solutions to CPUS utilization in iOS. Many apps, programmed to properly use these facilities, would benefit from more cores immediately. For most apps the sweet spot is around 4 cores these days before you see diminishing returns.
As well as separate chips to do the dirty work of all the other contextual interrupts (networking, motion,sound), it makes sense to focus on efficiency of fewer cores.
For a cell phone yes, the primary goal should be performance without killing thermals. You leave the cell phone world and things change dramatically as app performance becomes much more important. Separate extreme low power chips are a part of the equation but do realize that iOS still has many processes running in background which impacts CPUS usage. IOS may only support one foreground app but that isn't the only vide drawing power from the CPUs.
Perhaps the iPhone 6 has 2 cores while the iPad Air 2 has 4. Apple has a lot of code that can use multiple cores and they can really help when multi-tasking. More cores make it easier to drive data to the iPad's larger display.
That is a very good possibility. People need to realize that the app market for iPad is exploding, the only thing that holds the app world back is RAM and CPU performance.
Exactly- iOS doesn't need 4 cores. The only reason I could see them implementing it now is if legitimate multi-tasking is planning on coming to the iPad in the future and they want to make sure it has backwards capability.
Sorry Andy but this is baloney! IOS uses the same techniques as Mac OS to support multiple processors, any app that can benefit from cores on a Mac can likewise benefit from cores on an iOS device. I wouldn't be surprised to find iPhine sticking to dual core as power usage is paramount there, but on other platforms cores would be a huge benefit. More cores would enhance existing apps and allow new classes of apps to be delivered to iPads.
Again- That's a long shot. I just don't see the need
Where I do see the need with Apple's A7 (and subsequently A8 SoC) is some more RAM. 1.5-2gb (particularly on iPad) would go great lengths at expanding capabilities there. Anandtech mentioned the bottleneck occurs with RAM before CPU on the A7. The A8 will only lengthen that.
This I agree with 100%. Not just more RAM though it needs to be faster. I was actually a bit surprised to find that Apple did not debut an A7X variant as RAM speed was one of the things that "X" chips addressed. It is pretty obvious at times that current iOS devices are RAM starved.
All of that said- I can't wait to see whats released in September and October.
Exciting isn't it? No matter which side of the fence you are on a process shrink means new capabilities. It will be interesting to see how Apple addresses the potential of the new process.
In any event im unfortunately shocked by how readily some people accept the performance they are getting out of their iPads and iPhones. We have a long ways to go folks. I see a world where your phone docks with a desktop display morphing into a full fledged desk top machine. Drop said iPhone in your pocket and it would revert back to a simple iOS based iPhone.
Agree with you and Andysol. Yes, more cores can deliver nearly-linear performance gains (e.g. 12-core Mac Pro). <span style="line-height:1.4em;">This is mostly because of Apple's work on Grand Central Dispatch in both OS X (since 10.6) and iOS (since 4.0). GCD makes it vastly easier for developers to exploit multiprocessing, and optimizes it for them. Then there's OpenCL, which can use the massively parallel computing power of the GPU to perform general purpose computations if and </span>
when<span style="line-height:1.4em;"> the GPU has spare cycles.</span>
Nearly linear performance increases are rare in software. However a very large portion of the software base out there can benefit from more than 2 cores. I guess if Apple want to compromise they could offer a processor with three cores. However it makes more sense to split the line with two core chips going into the iPhones and four core chips going else where.
But multiprocessing (harnessing multiple CPUs) is useful at all times. Not just when you happen to be multitasking (which I assume you are defining as "switching between apps"). Even the most basic app needs multiple threads: o<span style="line-height:1.4em;">ne for handling GUI events and at least one background thread for doing computations to generate results to display. So any app can benefit from multiprocessing. Especially if if spawns many background threads. And it's easy to get many threads going, for example when fetching data from a URL and calculating a result to display and writing data to "disk" and updating iCloud all at once.</span>
It is surprisingly easy to keep all of those cores running for short periods of time, much much harder to leverage them for extended periods of time. However when you do leverage all the cores at once it has a dramatic impact on user perceptions of the machine.
So, of course, bringing up the subject of quad-core A8 SoCs in next-gen iOS devices might trigger battery life concerns. Well, my theory is that the faster you finish intensive processing tasks, the sooner you can revert to "idle." So your total power usage would increase logarithmically instead of linearly. Firing up 4 cores at once may draw more peak current, but you'd be drawing that current for a shorter time. I'm sure Apple can and will tweak performance vs. power usage in Grand Central Dispatch. And the larger frame of the (rumored) iPhone 6 model(s) might allow for a slightly larger battery. We'll see.
The surprising thing here is how little power an ARM takes. A few years ago, on much older processes the talk was that an ARM core could run at less that 500 milliwatts @ 2GHz. Of course that is just the core and not the supporting caches and other hardware. Often caches are the hottest running parts of a processor which is where you run into problems adding more cores.
<span style="line-height:1.4em;">Having said all that, I wouldn't be surprised if Apple did reserve the first quad-core A8 SoCs for the high-end of each iOS device line. Maybe they'll use the dual-core in the 4.7" iPhone and iPad mini Retina, and quad-core in the 5.5" iPhone and iPad Air. And maybe quad-core in a new high-end 4K Apple TV. Who knows?</span>
<span style="line-height:1.4em;">Oh, and I almost forgot about the software side of performance. At WWDC, Apple announced the Swift programming language and Metal graphics API, both next-gen technologies that Apple claims are vastly faster than legacy Objective-C and OpenGL. Looking forward to both.</span>
Who knows? Goid question - nobody on this forum obviously. However I could see a rational argument for splitting development across two chip lines. The more advanced chip could always be the chip going into iPhone "next" year after a process shrink. In other words a four core chip for this year's iPad would end up in iPhine next year after a process shrink.
I am rescinding my statement that Apple will not increase clock speed to 2 GHz.
Clock speed has nothing to do with big.LITTLE. For Apple it there are two considerations. The high priority one being power usage. The second is can the process node run their logic relaibly at those speeds.
Apple may increase clock speed to 2 GHz if they migrate to a big.LITTLE processor architecture.
All companies make mistakes and frankly big.LITTLE is perhaps a huge one on ARMs part. In any event Apples choice of clock speeds has nothing to do with big.LITTLE.
Qualcomm has announced ARM-based hexa-core and octo-core big.LITTLE 20 nm processor architectures at 2 GHz clock speeds using dual-channel 1600 MHz LPDDR4 RAM for early 2015.
I think this strengthens Dick Applebaum's argument.
Apple has so many other ways to go about saving power that I don't see any reason at all for them to do big.LITTLE. big.LITTLE just wastes die space for little in return. 2GHz shouldn't be seen as an unreasonable target to hit. Do some frequency scaling and the benefits of 2GHz can be balanced against power usage.
Then they wouldn’t have told him anything at all, so he shouldn’t be listened to.
Without more info that is an awfully strong statement to make! Before somebody like that would even consider signing up (agreeing to a NDA) he would have to have a conversation that is deep enough to allow a decision to be made. Like I said not everybody rolls over just because Apple comes knocking at the door.
Do you really believe Apple would have revealed anything specific about their intentions for Sapphire if he didn't agree to an NDA?
As far as I can see they didn't reveal anything to the professor. I suspect people are reading more into what the professor is saying here than was actually said. Being approached by someone doesn't mean you have had a extensive conversation with them.
As far as I can see they didn't reveal anything to the professor. I suspect people are reading more into what the professor is saying here than was actually said. Being approached by someone doesn't mean you have had a extensive conversation with them.
"I remember the Apple folk coming to speak to me about 18 months ago to discuss sapphire screens," Alford said.
I consider this a very important detail to reveal to anyone outside Apple that you're interested in using Sapphire as a screen (even though it might seem obvious at this point). Remember to this day Apple has not revealed their plans for the Sapphire produced by GTAT. If the professor said "Apple approached me about the possibility of using Sapphire in other areas of Apple devices" I would find it a bit more believable. He is claiming specifically that Apple mentioned screens which I don't think they would do without an NDA.
Without more info that is an awfully strong statement to make!
You’re saying he denied an NDA, which would imply Apple told him nothing, as there was no guarantee of secrecy. If he knows nothing, there’s no reason to listen to him.
If Apple did approach Neil Alford about Sapphire, that would make him a consultant. Apple would have compensated him and made him sign an NDA.
Really?
Prof Neil Alford, Head of Materials Science at Imperial College, expert in nano coating and dielectric loss, a long history of experimenting with Sapphire glass, gave this lecture: http://www.materials.ox.ac.uk/uploads/file/HRlecture2009.pdf
- all of which relate to the use and manufacture of touch enabled Sapphire screens.
I think it's fairly obvious why they consulted with him and published academic work is never subject to NDAs if all you want to do is tap someone's theoretical knowledge on the subject.
Who cares what anyone's opinion was 18 months ago on Sapphire glass. You don't think Apple has improved the technology by leaps and bounds by then? I'm sure the performance/price has improved significantly. We shall see of course. My money is on Apple.
Yes I would have to agree and any comments he is now making about Apple and it use of Sapphire would be a breach of that NDA.
Not necessarily. It would depend on what he was being consulted about. I've signed more than a few NDA's, and had others sign them. Not everything you talk about requires an NDA. If Apple asked him general questions about saphirre, something about which he is a known expert, there is no reason that I can think of offhand that would require him to sign an NDA.
I've heard that the new iPad will have 512 MB of RAM, down from 1 GB. Apparently, most people only have one or two tabs open in Safari, so don't need the extra.
Comments
Then they wouldn’t have told him anything at all, so he shouldn’t be listened to.
A performance enhanced A8 would be good enough for iPhone, however it won't do for iPad and some other things Apple could be working on. This is why I've stated before that I could see "X" variants again of Apples processors. With the X variants targeting higher performance machines, it would be easy for Apple to deliver a highly optimized cell phone chip.
Beyond that I've heard solid rumors that Apples goals this go around have been improved power efficiency. People might end up underwhelmed by the next iPhone processor if they don't value battery life.
They could have contacted him about an NDA and he could have said no. Not everybody rolls over when Apple comes knocking.
If A8 is dual core, good. Glad Apple isn't jumping on the "more cores for the hell of it" bandwagon. They know what theyre doing when it comes to chip design, which is why the A7 still blows everything else out of the water in most respects, even Samsung's 8 core mobile chips.
Whatever Apple do with the chip, they make the most out of it. Samsung's octa-core chip was nothing more than just a gimmick because it only worked 4 cores at the time. How pathetic! That's why Samsung decided to use SnapDragon quad-core chip in their GS4 handset for US market instead of their own Exynos Octa-core (only for European GS4).
Man, it’s a good thing you don’t actually believe this; it’s one of the best pieces of trolling I’ve seen in a while.
That would be great for iPhone, maybe not so great for iPad, an enhanced AppleTV or other devices Apple has coming.
Enough for a cell phone yes, maybe not for an iPad or enhanced Apple TV.
As it is im not sure you have a strong understanding of Grand Central Dispatch and other solutions to CPUS utilization in iOS. Many apps, programmed to properly use these facilities, would benefit from more cores immediately. For most apps the sweet spot is around 4 cores these days before you see diminishing returns.
For a cell phone yes, the primary goal should be performance without killing thermals. You leave the cell phone world and things change dramatically as app performance becomes much more important. Separate extreme low power chips are a part of the equation but do realize that iOS still has many processes running in background which impacts CPUS usage. IOS may only support one foreground app but that isn't the only vide drawing power from the CPUs.
That is a very good possibility. People need to realize that the app market for iPad is exploding, the only thing that holds the app world back is RAM and CPU performance.
Exciting isn't it? No matter which side of the fence you are on a process shrink means new capabilities. It will be interesting to see how Apple addresses the potential of the new process.
In any event im unfortunately shocked by how readily some people accept the performance they are getting out of their iPads and iPhones. We have a long ways to go folks. I see a world where your phone docks with a desktop display morphing into a full fledged desk top machine. Drop said iPhone in your pocket and it would revert back to a simple iOS based iPhone.
Who knows? Goid question - nobody on this forum obviously. However I could see a rational argument for splitting development across two chip lines. The more advanced chip could always be the chip going into iPhone "next" year after a process shrink. In other words a four core chip for this year's iPad would end up in iPhine next year after a process shrink.
Apple has so many other ways to go about saving power that I don't see any reason at all for them to do big.LITTLE. big.LITTLE just wastes die space for little in return. 2GHz shouldn't be seen as an unreasonable target to hit. Do some frequency scaling and the benefits of 2GHz can be balanced against power usage.
Without more info that is an awfully strong statement to make! Before somebody like that would even consider signing up (agreeing to a NDA) he would have to have a conversation that is deep enough to allow a decision to be made. Like I said not everybody rolls over just because Apple comes knocking at the door.
As far as I can see they didn't reveal anything to the professor. I suspect people are reading more into what the professor is saying here than was actually said. Being approached by someone doesn't mean you have had a extensive conversation with them.
As far as I can see they didn't reveal anything to the professor. I suspect people are reading more into what the professor is saying here than was actually said. Being approached by someone doesn't mean you have had a extensive conversation with them.
"I remember the Apple folk coming to speak to me about 18 months ago to discuss sapphire screens," Alford said.
I consider this a very important detail to reveal to anyone outside Apple that you're interested in using Sapphire as a screen (even though it might seem obvious at this point). Remember to this day Apple has not revealed their plans for the Sapphire produced by GTAT. If the professor said "Apple approached me about the possibility of using Sapphire in other areas of Apple devices" I would find it a bit more believable. He is claiming specifically that Apple mentioned screens which I don't think they would do without an NDA.
You’re saying he denied an NDA, which would imply Apple told him nothing, as there was no guarantee of secrecy. If he knows nothing, there’s no reason to listen to him.
Really?
Prof Neil Alford, Head of Materials Science at Imperial College, expert in nano coating and dielectric loss, a long history of experimenting with Sapphire glass, gave this lecture:
http://www.materials.ox.ac.uk/uploads/file/HRlecture2009.pdf
- all of which relate to the use and manufacture of touch enabled Sapphire screens.
I think it's fairly obvious why they consulted with him and published academic work is never subject to NDAs if all you want to do is tap someone's theoretical knowledge on the subject.
It was a good lecture.
Man, it’s a good thing you don’t actually believe this; it’s one of the best pieces of trolling I’ve seen in a while.
Thank you!
I'm very glad you caught my humor.
Who cares what anyone's opinion was 18 months ago on Sapphire glass. You don't think Apple has improved the technology by leaps and bounds by then? I'm sure the performance/price has improved significantly. We shall see of course. My money is on Apple.
Not necessarily. It would depend on what he was being consulted about. I've signed more than a few NDA's, and had others sign them. Not everything you talk about requires an NDA. If Apple asked him general questions about saphirre, something about which he is a known expert, there is no reason that I can think of offhand that would require him to sign an NDA.
Sounds good to me.