Apple hogging Toshiba memory; future 1GHz iPhone chip?

[AppleInsider]

Apple's demand for flash memory is proving insatiable in the run-up to new iPods, according to one report. Also, Samsung has developed technology that could increase the processor speed of future Apple handhelds.

Toshiba may be overwhelmed by Apple memory orders

Mixed messages from Toshiba may be a clue to a major order of NAND flash memory from Apple, if claimed sources at circuit makers are accurate.

The apparent insiders for DigiTimes say Toshiba is boosting its production to a high 90 percent of its capacity next month but, strangely, is telling customers in the spot market -- companies that buy on short notice -- that it won't have much supply for their orders. The combination is usually a sign that a long-term client is swallowing up most of the components and leaving little else for the smaller companies that can't always make these deals.

It's speculated that the mystery drain on flash memory is none other than Apple, which has a history of creating shortages in the flash memory market whenever it's gearing up for the release of a new iPhone or iPod. The American firm also recently confirmed a $500 million contract with Toshiba that will guarantee a healthy supply of NAND flash for an unspecified amount of time, making it the most probable source of the problem.

Like most home electronics companies, Apple normally queues up production weeks or months ahead of when it actually intends to ship its products and is more than likely bracing itself for holiday sales of new mobile devices that could include iPods with cameras.

Samsung details 1GHz ARM processor

On Tuesday, Samsung escalated the race to faster mobile processors with word of its first ARM-based processor design to be built on a 45 nanometer assembly process.

So far known only as Hummingbird, it would use the smaller, cooler running architecture to increase the maximum clock speed of future system-on-chip processors to 1GHz, or significantly above Samsung's current 833MHz peak, without consuming more energy or wasting more heat. It would use the same Cortex-A8 platform found in the iPhone 3GS and would be Samsung's fastest ARM processor as a result.

Whether or not it will ever reach an Apple mobile device is up in the air, however. While every iPhone and iPod touch to date has used a Samsung-designed chip, Apple has signaled its intent to use custom-designed processors from recent acquisition P.A. Semi sometime in the future, avoiding any dependency on Samsung's plans for improving its own mobile hardware.

[lightstriker]

At least non of this shortage has anything to do with the iTablet phantom.

[melgross]

Survival of the fittest, or most popular. Apple is doing what they should be doing here, insuring a ready flow of parts at lower prices.

A 1GHz processor sounds interesting. If this tablet thing is real, it will need something more powerful than the one in the iPhone. It might also need something more powerful than the Atom 1.67 GHz chip.

I have to say that the Toshiba mini NB205 netbook my daughter is using this summer is slow. Even with the upgrade to 2 GB RAM, it's slow.

I can't imagine a tablet from Apple being this slow. But the Atom is supposed to be more powerful than the fastest ARM.

How will this work? A large device could have a pretty big battery, so maybe two chips and the more powerful graphics chip? That would do it.

How about a 9.7" OLED? Sigh! That would be nice, but the rumors of screen sales to Apple hasn't mentioned that possibility.

[melgross]

Originally Posted by lightstriker

At least non of this shortage has anything to do with the iTablet phantom.

Maybe, maybe not.

[Hattig]

I presume the Samsung SoC is a die-shrink of their current offering. Could be that Apple will switch silently to this without changing operating parameters. Maybe Jan'10 iPhone 3GS' will get an extra hour of battery.

But yes, Samsung's actual performance with their ARM SoCs does make Apple's PA Semi purchase seem worthless, so PA Semi must be designing other chips for Apple beyond what Apple would have otherwise got from Samsung.

[Hattig]

Originally Posted by melgross

I have to say that the Toshiba mini NB205 netbook my daughter is using this summer is slow. Even with the upgrade to 2 GB RAM, it's slow.

Yeah, well Intel's current netbook offering is a CPU + dumb chipset, in effect.

ARM is more about dedicated hardware for functionality, so it had video decode, hardware encryption and so on. The CPU might be slower overall (but they're so small that you can have several of them) but the overall effect is good.

You'll notice if YouTube playback skips, but can cope with something taking a little longer to occur on a once-off basis. The ARM platform solves the common case with dedicated hardware, whilst using 1/10th of the power overall.

Look at Intel's Larrabee - they've had to add two hardware video decoders to the design because it uses too much power to do it in software on the general purpose hardware. To be honest, this says all you need to know about Larrabee version 1. Just ignore it. And version 2. Version 3 could be good though, many companies get version 3 right...

[Shrike]

The rumors say that PA Semi was split into 2 teams. One team designing for the "iTablet" and the other for a next-gen iPhone/iPod. If the supposed Apple tablet comes out, I predict it will be a dual-core Cortex A8 at 833 MHz and dual Power SGX GPUs all in one package with 1 GB memory and 64 GB flash storage. There was a rumor awhile back that iPhone OS X 3.x was "quad" capable.

Your guess is as good as mine on what "quad" meant. It'll be fast enough. It won't support Adobe Flash nor Java.

[PaperMacWriter]

Samsungs 1GHz ARM can only be good news, as there are only four possible options I see Apple can take, all being good/neutral:

1) Apple uses this in iPod Touch and 3GS at full speed or with minor under-clocking (Good, but unlikely)

2) Apple uses at 500MHz in 3GS and iPod Touch (good, fairly likely)

3) Apple doesn't use in favor of P.A. Semi ARMs (Good, unknown)

4) Apple ignores this completely (neutral, semi- likely)

As for the Toshiba stuff, that could be more interesting news if we knew wht size chips they were making. Say, 64GB Whatever's being cooked up in Cupertino, I'm sure it'll amaze us

SG

[melgross]

Originally Posted by Hattig

I presume the Samsung SoC is a die-shrink of their current offering. Could be that Apple will switch silently to this without changing operating parameters. Maybe Jan'10 iPhone 3GS' will get an extra hour of battery.

But yes, Samsung's actual performance with their ARM SoCs does make Apple's PA Semi purchase seem worthless, so PA Semi must be designing other chips for Apple beyond what Apple would have otherwise got from Samsung.

If the chip runs at 1 GHz, then battery life would be about the same. That's a big boost in performance. I don't see Apple moving the iPhone to this now. It's an appliance. Like a game machine. With a shorter lifespan.

Likely, PA will work with Samsung to produce their chips, as they aren't a fab. Any advantage Samsung has here would be used by PA as well.

[Logisticaldron]

Originally Posted by lightstriker

At least non of this shortage has anything to do with the iTablet phantom.

I wouldn’t say that. The suggested tablet device seems more of a complimentary device for your main computer, not a replacement for it, so expecting it to use a HDD is not where I’d put my money. Apple has almost completely done away with the 1.8” HDDs and a 2.5” HDD would probably be too large for the device, especially if they want it to be thin and light. I’d think that a 1.8” SSD or just a bunch of Flash on the system board would be the way Apple would go. Just my theory.

Originally Posted by melgross

I have to say that the Toshiba mini NB205 netbook my daughter is using this summer is slow. Even with the upgrade to 2 GB RAM, it's slow.

I can't imagine a tablet from Apple being this slow. But the Atom is supposed to be more powerful than the fastest ARM.

I assume it’s running XP. Tests show that OS X is faster on the slower Atoms and are much better with battery management. Besides that, having a full version of Mac OS X may not be right for this class device. For a tablet to work I think that a new UI and some iPhone-like frameworks may have to be developed. A 3rd branch for OS X between the Mac and the iPhone/Touch. After making Mac OS X work so well on the iPhone with an ARM CPU it seems like a given that Apple couldn’t make a hybrid version of OS X for a hybrid-like device.

[SpamSandwich]

"iTablet", no... The all new reimagined iBook, yes.

[TenoBell]

Originally Posted by melgross

I have to say that the Toshiba mini NB205 netbook my daughter is using this summer is slow. Even with the upgrade to 2 GB RAM, it's slow.
I can't imagine a tablet from Apple being this slow. But the Atom is supposed to be more powerful than the fastest ARM.

Does she have Windows or did you hackintosh it?

How about a 9.7" OLED? Sigh! That would be nice, but the rumors of screen sales to Apple hasn't mentioned that possibility.

There is a history of introducing newer hardware in the Touch, then bribing it to the iPhone in the next revision. The Touch may get OLED first then the iPhone next year. There isn't any reason for Apple to delay.

[Cubert]

The title for this article could have easily been, "Apple Covering All Bases".

[MsNly]

Originally Posted by SpamSandwich

"iTablet", no... The all new reimagined iBook, yes.

I always loved my iBook

[Ronbo]

Originally Posted by melgross

If this tablet thing is real, it will need something more powerful than the one in the iPhone. It might also need something more powerful than the Atom 1.67 GHz chip.

(Seriously) Is that expectation based on what you expect the tablet to be able to do, or is it based on an idea that even an iPod Touch with a big screen would require a larger processor? Perhaps the intended function of the thing is such that all it needs to better video hardware.

(Less seriously) It sometimes seems as though talking about iTablet specs is akin to talking about the capabilities of the Flying Dutchman. One gets the feeling that if the iTablet ever sails into port, most everyone will be angry and disappointed, loudly decrying its lack of capabilities, expectations of which it could only have satisfied with supernatural help

If it is based on the iPhone OS, it might not need anything much more powerful than what's in the current machines. I remember the amazing things my Newton could do, and that was a wind-up toy compared with current technology. If the iTablet is real and not just an idle skunkworks project, and if it really does need a powerful processor, then it might be awaiting processor. Perhaps the signal people should be looking for that signals the coming of the iTablet is a more powerful processor. If you don't believe the Cortex is fast enough, when is the next generation?

[GregAlexander]

Originally Posted by Hattig

ARM is more about dedicated hardware for functionality, so it had video decode, hardware encryption and so on. The CPU might be slower overall (but they're so small that you can have several of them) but the overall effect is good.

I agree.

I'm kind of wondering whether we'll see ARM in the next AppleTV, actually.

[OC4Theo]

For owners of iPhone 3G, here is another reason to wait till next summer to get a new iPhone.

I was tempted to rush out and buy the new iPhone 3GS last month, but I decided to wait. For people with deep pockets, it's not a big deal. Just got mine in September after using the original iPhone since 2 weeks of its original launch.

Next summer, a faster, better iPhone will be introduced. That is for sure. That is when I will get a new one. Who knows, it might even be cheaper. For now 3G will do.

[Porchland]

Originally Posted by lightstriker

At least non of this shortage has anything to do with the iTablet phantom.

Hide and watch.

[L255J]

I personally like how "iNote" sounds. Maybe "Pad".

[melgross]

Originally Posted by Logisticaldron

I assume it’s running XP. Tests show that OS X is faster on the slower Atoms and are much better with battery management. Besides that, having a full version of Mac OS X may not be right for this class device. For a tablet to work I think that a new UI and some iPhone-like frameworks may have to be developed. A 3rd branch for OS X between the Mac and the iPhone/Touch. After making Mac OS X work so well on the iPhone with an ARM CPU it seems like a given that Apple couldn’t make a hybrid version of OS X for a hybrid-like device.

I would NOT like to see Apple branch off the OS once again. not at all.

What I would like to see, and it might not be too difficult to do is to have both the Simple Finder from OS X on the machine as default, then allow the OS to also use an iPhone GUI as well, so that it could run both types of apps.

But that would b a problem with either cpu. It perhaps would need both. Unless there was a way to run iPhone apps in emulation without losing speed.

This would be a very good convergence device.

I'm not sure apple would want such a thing though.

Hey, Ireland, where are you when we need you?

[melgross]

Originally Posted by TenoBell

Does she have Windows or did you hackintosh it?

She's using XP. You know my views about copyright and eula's.

But, we spoke to her today.

Ah, Windows, ya gotta love it.

She doesn't like the trackpad, so she bought a mouse. She plugs it in to install it. The machine says the usual junk about new hardware, and so she installs it. It says it's fine, and she can use the new mouse. But it doesn't work!

She uninstalled it and re-installed it three times. Still didn't work. When we spoke I told her a couple things to check, which she did, and it STILL didn't work.

She says she hates Windows. We're going to break down and get her the 13" MacBook Pro next week.

When I left her, she was still trying to figure why it won't install.

There is a history of introducing newer hardware in the Touch, then bribing it to the iPhone in the next revision. The Touch may get OLED first then the iPhone next year. There isn't any reason for Apple to delay.

Isn't bribery illegal?

Well, I'd like to see OLED, sevral companies have already gone to it.

But I wonder if those companies made their phones to fit the available screens, rather than the other way around.

Apple won't do that, I don't think. A screen that will fit the form factor will be needed. With Apple's volumes, it may happen.

You know, a slightly larger OLED screen can fit into the current cases, as OLED's need less edge around the screen proper than LCD's do. We could get a screen that's up to .25" wider, and .375" longer.

I don't know if that would happen.

[ksec]

The problem is Snapdragon manage 1.3Ghz at 45nm, alothough drawing slightly more power, Snapdragon already includes 3G , WiFi and Bluetooth.

Although not proved by benchmarks, Qualcomm claims their SnapDragon is clock to clock faster then other Cortex A8 implmentation.

[Logisticaldron]

Originally Posted by melgross

I would NOT like to see Apple branch off the OS once again. not at all.

Unless they include a capacitance stylus I don’t see how they can’t alter the UI to fit a large screen device that works primarily with one’s fingers/

It perhaps would need both. Unless there was a way to run iPhone apps in emulation without losing speed.

This would be a very good convergence device.

I'm not sure apple would want such a thing though.

That is the easy part because it’s already been done to some extent. The iPhone Simulator in the SDK works very well. There are caveats with it running on x86 with a different GPU, but on modern Mac notebooks it’s considerably faster than on the iPhone.

[melgross]

Originally Posted by Ronbo

(Seriously) Is that expectation based on what you expect the tablet to be able to do, or is it based on an idea that even an iPod Touch with a big screen would require a larger processor? Perhaps the intended function of the thing is such that all it needs to better video hardware.

(Less seriously) It sometimes seems as though talking about iTablet specs is akin to talking about the capabilities of the Flying Dutchman. One gets the feeling that if the iTablet ever sails into port, most everyone will be angry and disappointed, loudly decrying its lack of capabilities, expectations of which it could only have satisfied with supernatural help

If it is based on the iPhone OS, it might not need anything much more powerful than what's in the current machines. I remember the amazing things my Newton could do, and that was a wind-up toy compared with current technology. If the iTablet is real and not just an idle skunkworks project, and if it really does need a powerful processor, then it might be awaiting processor. Perhaps the signal people should be looking for that signals the coming of the iTablet is a more powerful processor. If you don't believe the Cortex is fast enough, when is the next generation?

It's a bit of both, I think.

Wouldn't you expect a much larger device to do more?

That's why so many people are disappointed with netbooks. They think they're just smaller, and cheaper notebooks. But they're not. They're REALLY slow.

Would Apple allow that? I don't think so. Managing expectations will be very important here.

But people expect things to respond at a certain rate. With a phone, people have been accustomed to crappy performance, or clumsy usage, so the 2G and 3G's performance wasn't too much of an issue. But the new 3Gs is much faster. I saw that with my daughter's new one. Really noticeable.

With such a large screen, the graphics performance will be really slow unless Apple uses the fastest graphics chip for the ARM that they can get from Imagination, the PowerVR SGx543 at 35 Million polygons/sec, and 16 instances. Right now, the one they use, I think the number is, the 540 at 14 million, and 4 instances.

But would one ARM, even if it's 1GHz be enough? The 2 core version wont be out for some time.

It wouldn't be good if the 3GS outperformed the tablet.

[melgross]

Originally Posted by Logisticaldron

Unless they include a capacitance stylus I don’t see how they can’t alter the UI to fit a large screen device that works primarily with one’s fingers/

It would be interesting, wouldn't it?

I was skeptical at first about the iPhone. I was one who was asking, and experimenting with a stylus, but I got over it. It's performed much better than I thought.

I think it would work with a much bigger screen. The multitouch already works very well. Even on my iPhone, I can touch a really small link between two other links, and almost all the time, the correct link is chosen. That's without magnifying the screen first. Amazing! I've tried it on some other phones and it doesn't work as well. I tried the Palm, but the screen is even smaller, and the links are almost impossible to see, unless you get real close, and the smaller screen makes it more difficult to correctly jab the correct one.

The thing about multitouch is that it senses a fair number of spots around where you hit, and uses an algorithm to figure out where you were selecting. It works very well. A larger screen should allow this to function pretty accurately I would think.

Also, this is a big device, Apple could bless us with a couple of ports, a mini Displayport, and a USB 2. That would allow a mouse, or track ball (or gamestick!). A KEYBOARD!!!

If the mini DP port wasn't invented for this, then what's it good for?

That is the easy part because it’s already been done to some extent. The iPhone Simulator in the SDK works very well. There are caveats with it running on x86 with a different GPU, but on modern Mac notebooks it’s considerably faster than on the iPhone.

The iPhone simulator is working on MUCH more powerful processors. The tablet, if it had an Atom, say, wouldn't have that processing power.

[TenoBell]

Originally Posted by melgross

She's using XP. You know my views about copyright and eula's.

Yes I know, I would have been shocked if you hacked it.

Isn't bribery illegal?

Well, I'd like to see OLED, sevral companies have already gone to it.

My copy editor didn't catch the typo.

Apple seems to like to offer better hardware in the refresh of the Touch that was not offered in the previous refresh of the iPhone. Seems like the most logical way they would introduce OLED.

[melgross]

Originally Posted by TenoBell

Yes I know, I would have been shocked if you hacked it.




My copy editor didn't catch the typo.

Apple seems to like to offer better hardware in the refresh of the Touch that was not offered in the previous refresh of the iPhone. Seems like the most logical way they would introduce OLED.

I'd like that, but what would it do to iPhone sales? With the camera and mike, an OLED could have serious effects.

[Logisticaldron]

Originally Posted by TenoBell

Apple seems to like to offer better hardware in the refresh of the Touch that was not offered in the previous refresh of the iPhone. Seems like the most logical way they would introduce OLED.

This would be the way to offer a higher resolution display, too, this way developers that haven’t been building their apps using vector graphic images or skirting Apple’s other development guidelines will have nearly a year to get it right for the next iPhone, which is much more important to users to work right than an iPod, for obvious reasons.

It might not come this year, but eventually a higher pixel per inch will be needed, if only to keep up with the pack.

[GregAlexander]

Originally Posted by melgross

What I would like to see, and it might not be too difficult to do is to have both the Simple Finder from OS X on the machine as default, then allow the OS to also use an iPhone GUI as well, so that it could run both types of apps.

The thing is, counter to rational thought , if the tablet offers MORE functionality it might cause it to fail.

We know the history of tablets - they haven't been particularly well received. If Apple brings out a tablet that is capable of running OSX apps, then we get access all the applications on OSX today - applications that aren't designed with virtual touch interface and virtual keyboard etc in mind.

If we are using many of those apps, then we start to naturally wish we had a keyboard and mouse. And eventually we would rather just have a laptop. Additionally, the extra processing power required for something designed to run OSX apps means it has to cost something like a laptop, and have something like the thickness of a laptop (and, like MBA, if it's extra thin that will result in extra cost too).

I say force developers to write new apps for it, current OSX apps shouldn't run. I'm in 2 minds about making it capable of running standard iPhone apps - I'll swing towards the same though - BLOCK that, BUT make it easy to modify iPhoneOS apps for the tablet.

If a tablet is to succeed it has to break the mold and our expectations.

[GavinScrimgeour]

I reckon your expectations of the Apple iTeamaker with it's mouth and leg attachments for it's all singing and dancing capabilities are well over estimated.

As Apple have said cheerio to the hard disk iPod i.e. iPod classic, the flash memory will be bound for the Classic as a replacement storage medium.

We have to remember that we're still in the recession and it would be highly unlikely Apple would turn around and offer such a revolutionary product in times of woe.

As the Classic is the only iPod still using a hard disk for storage, we can only assume it will be replaced by flash memory bought by manufacturers who have bought a lorry load of it in at bulk discount, to be used in an iPod to sell at a "reasonable" price to the consumer.

I'd reckon an 64GB iPod Classic with the Flash memory would fetch about £200.

[ArthurAscii]

Originally Posted by GregAlexander

I'm kind of wondering whether we'll see ARM in the next AppleTV, actually.

ARM's advantage for portables - low power consumption at a given level of performance - would not be relevant here, although ARM does show up in routers and so forth. They and their content-providers might be tempted by the lockdown that custom hardware could offer for both iTablet and AppleTV. But I'd worry if Apple started to custom-design ARM chips right across their product range. They could end up spending a great deal of money for no better than industry-average performance.

[GregAlexander]

Originally Posted by ArthurAscii

ARM's advantage for portables - low power consumption at a given level of performance - would not be relevant here, although ARM does show up in routers and so forth. They and their content-providers might be tempted by the lockdown that custom hardware could offer for both iTablet and AppleTV.

I agree the lower power consumption is not an advantage in the AppleTV.
I thought ARM chips were cheaper than Intel chips though - perhaps I'm wrong.

But I'd worry if Apple started to custom-design ARM chips right across their product range. They could end up spending a great deal of money for no better than industry-average performance.

That's true. Right now though the existing iPod Touch internals seem to be enough to drive an SD version of the AppleTV... just remove the battery & screen, add an IR port. PERHAPS replace the solid state drive with a hard disk? The prices of entry-level iPod Touch and AppleTV are the same at present.

If the next iPod Touch chip can handle HD video (ie NOT a custom-designed ARM chip), is there any advantage to using Intel in the AppleTV?

[wizard69]

Originally Posted by Hattig

I presume the Samsung SoC is a die-shrink of their current offering.

it is hard to say for sure but I'd say it is not likely. The reason being is that Samsungs current offerring is A8 and it is likely that new designs are A9 Cortex based.

Could be that Apple will switch silently to this without changing operating parameters. Maybe Jan'10 iPhone 3GS' will get an extra hour of battery.

Well that would be nice but do you really think Apple would be that generous? Personally I can see Apple reving the iPhone before it's scheduled time.

But yes, Samsung's actual performance with their ARM SoCs does make Apple's PA Semi purchase seem worthless, so PA Semi must be designing other chips for Apple beyond what Apple would have otherwise got from Samsung.

Apparently you don't understand this industry. First; PA Semi is a design house not a manufacture. They still need a foundry to produce the chips. Second; this means there is a production ready 45nm ARM chip which is a big milestone. Third; if Samsung is willing to fab a full custom SoC for Apple/PA Semi, it highlights the possibility that the SoC will be 45nm and very cool running. If Apple implements the best ARM has to offer, that is a quad Cortex SoC, we could have some very snappy machines.

As to PA I actually tend to agree that that they are designing for more advance things than a simple iPhone. Base smart phones are commodities and need cheap but powerful processors, that means very high volume commodity devices. So I believe you are right in part, PA Semi is likely doing more design work than Apples has acknowledged. This should not surprise anybody though.

All in all I think this annoucement highlights that Apple could have some very bleeding edge products in the pipeline. This is not a negative revelation at all.



Dave

[wizard69]

Originally Posted by Logisticaldron

It might not come this year, but eventually a higher pixel per inch will be needed, if only to keep up with the pack.

Actually it is the last thing needed. It makes no sense to have the pixels so small that you need a microscope to see them. IPhone & Touch are doing fine right now. What is needed is a bigger screen for more pixels.

Dave

[mjtomlin]

Originally Posted by Hattig

But yes, Samsung's actual performance with their ARM SoCs does make Apple's PA Semi purchase seem worthless, so PA Semi must be designing other chips for Apple beyond what Apple would have otherwise got from Samsung.

Seem worthless? How exactly? The size and relevant speed of a chip doesn't have much to do with the actual design. Both of those are results of manufacturing processes, not chip design.

SoC's contain many different components all thrown onto one chip. It's basically a system on a chip. By designing their own, Apple can customize it anyway they wanted. They could potentially put core OS X functionality directly on the chip. There are many different things they could do to customize the SoC and give them an advantage the competition would never be able to duplicate.

[ArthurAscii]

Originally Posted by GregAlexander

I thought ARM chips were cheaper than Intel chips though - perhaps I'm wrong.

I don't know the answer to this. Perhaps if Apple can get sufficient volume, as they clearly can with iPhone/iPod components (a very different device), it would be cost-effective. But Apple TV currently uses a 1Ghz Pentium M, so the choice would be commodity x86 versus bleeding-edge ARM with upfront design costs. People do say Apple TV runs awfully hot - can anyone confirm this?

The custom hardware inside games consoles, set-top boxes and DVRs is, I believe, usually subsidised over the life of the device with income from the software, which has not been Apple's model, although with the iTunes ecosystem that is becoming an option - and I'd expect to see it on Apple TV first.

[Hattig]

Originally Posted by GregAlexander

I agree.

I'm kind of wondering whether we'll see ARM in the next AppleTV, actually.

I'm actually surprised that Apple hasn't done this already. The iPhone 3GS SoC is more than enough to handle what the AppleTV does, and this 45nm shrink/enhancement even more so. I can only imagine that Apple is waiting on PA Semi to deliver.

Maybe the tablet will be the first device that uses PA Semi technology, in which case I am hoping for a dual-core Cortex A9 with PowerVR SGX543 graphics and video acceleration. However I think we're a year off PA Semi hardware being available in consumer devices.

[mjtomlin]

Originally Posted by ArthurAscii

I don't know the answer to this. Perhaps if Apple can get sufficient volume, as they clearly can with iPhone/iPod components (a very different device), it would be cost-effective. But Apple TV currently uses a 1Ghz Pentium M, so the choice would be commodity x86 versus bleeding-edge ARM with upfront design costs. People do say Apple TV runs awfully hot - can anyone confirm this?.

It does run hot, not sure what that's from though. I suppose we can just blame the Intel chip and be done with it, although, I know I NEVER put my AppleTV to sleep, so it's constantly on and running. I suppose I should put it to sleep every night when I'm done with it. Oh well.


Anyway, I think Apple will eventually switch to an ARM CPU in the AppleTV. There's no reason for it to have Intel compatibility, as it doesn't run full Mac OS X due to its limited interface, nor is it conducive to running desktop based applications. I'm guessing they are going to release two products in the near future, both of them running the same ARM CPU. One will be a new AppleTV and the other will be some kind of touch screen based portable that will redefine portable computing... combining both multi-touch and a more traditional pointing device interface. I don't believe it will be a shrunken Mac OS X, or an enlarged iPhone OS. It will still be OS X, just another version made for a specialized purpose. It will be a device that sits between Macs and iPhones, in form, function and price.

[Hattig]

Originally Posted by wizard69

it is hard to say for sure but I'd say it is not likely. The reason being is that Samsungs current offerring is A8 and it is likely that new designs are A9 Cortex based.

Unlikely. A die shrink with a current product is a very common practice when migrating to a new process. Also the A9 designs were meant to be coming to market next year, I'm not sure if Samsung would have been able to integrate it and get it into production, on a new process no less, this far ahead of anyone else.

Apparently you don't understand this industry. First; PA Semi is a design house not a manufacture. They still need a foundry to produce the chips. Second; this means there is a production ready 45nm ARM chip which is a big milestone. Third; if Samsung is willing to fab a full custom SoC for Apple/PA Semi, it highlights the possibility that the SoC will be 45nm and very cool running. If Apple implements the best ARM has to offer, that is a quad Cortex SoC, we could have some very snappy machines.

I think you misread what I wrote. Anyway, Samsung are currently acting as design house and foundry, and performing very well, and selling tens of millions of these chips to Apple. Oddly enough, Samsung's phone division uses TI OMAP chips... :S

There's no reason for Apple to continue using Samsung as a foundry with their own designs, they could use TSMC, Global Foundries, TI, UMC, etc, instead.

[Hattig]

Originally Posted by mjtomlin

Seem worthless? How exactly? The size and relevant speed of a chip doesn't have much to do with the actual design. Both of those are results of manufacturing processes, not chip design.

SoC's contain many different components all thrown onto one chip. It's basically a system on a chip. By designing their own, Apple can customize it anyway they wanted. They could potentially put core OS X functionality directly on the chip. There are many different things they could do to customize the SoC and give them an advantage the competition would never be able to duplicate.

Maybe I have to break my point down into a list.

1) Samsung are perfoming well in terms of ARM SoCs.

2) Therefore no need to Apple to just use PA Semi for iPhone SoCs, because Samsung are providing good SoCs themselves for phone use, for good prices (according to iSupply teardowns anyway, the SoC price is fairly low). And Samsung would be more than happy to customise the design for Apple.

3) Therefore PA Semi is definitely designing more complex chips.

4) Otherwise why buy PA Semi? Upfront design costs for a SoC are massive, and that's on top of the PA Semi purchase price. Each SoC design will end up in around 40m Apple devices (including iPod nano here), so it will be hard for Apple to get the cost per chip down and competitive with what Samsung offer. OTOH it was pocket change for Apple, and yes, it does mean that Apple can include custom hardware to differentiate (or more likely, lock down) their product. Presumably Apple's high-end 45nm SoC (Apple TV / Tablet) will become the mass-market 32nm SoC for iPhone a year later, thus extending the life of a design. Thus I think that PA Semi's main project is the high-end SoC - dual or quad A9, etc, etc, we'll find out one day.