Slashdot Mirror
Intel To Build Next Gen Processor For iOS Devices
BogenDorpher writes "It looks like Apple will be using Intel as a main processor manufacturer to power the iPad, iPod touch, and the iPhone. Apple, who currently uses Samsung, will focus on making a switch to Intel within a year."
Apple hasn't been happy with Samsung launching android phones, and this is how they're showing their displeasure.
I wonder if they has anything to do with the Samsung and Apply suing each other
I thought Intel only did x86/64 and Samsung didn't do either. Is this another PowerPC->Intel type move from Apple or am I missing something (quite likely)?
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
Comment removed based on user account deletion
Meanwhile MS has just started chasing ARM.
And by "just started" you mean they've had versions of Windows on ARM for going on near 15 years?
The report from EETimes suggests Intel is only going after foundry business to produce the A-series processors for Apple, not that Apple is looking to change architectures.
It could be Apple leaving Samsung, or it could be they've decided to go with multiple suppliers for everything to reduce potential impacts from future disasters.
When Apple switched to Intel chips a few years back, I remembered all the venom spewed toward Intel by all my Apple-obsessed friends over the previous 20 years.
Now they cherish their Intel chips. But they still bash MS. Why, I got an Outlook e-mail from one of my Apple friends just yesterday, sending me a Powerpoint presentation he had made on his Mac, with a funny joke about how lame MS is.
I had no problem opening it in OpenOffice on my AMD-powered CentOS box.
Probably because (quoting Wikipedia): "P. A. Semi (originally "Palo Alto Semiconductor"[1]) was a fabless semiconductor company"
You still need a fab. Apple already knows how to design CPUs.
It appears from casual googling, that Intel could make the A5 using a smaller process size than the current ARM manufacturers are able to produce.
If you mod me down, I shall become more powerful than you could possibly imagine.
They did keep the bit that they use for their RAID cards; but the general-purpose processor side went to Marvell.
It would be quite interesting, though, if this was a case of Intel taking a contract fab job. Traditionally, they haven't done that(at least with their leading edge process stuff, I don't know what they do with older fabs). Intel doing an apple-exclusive run of ARM chips on the same process they do their x86s on would be dramatic and probably make a bunch of people rather sad pandas...
Here's a similar report from EETimes.
Of course that article says that the "Next Gen Processor For iOS Devices" (as well as the current A5) will be build by Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC) (at least some of them), and that Intel may want to build the Gen after that.
Fandroids hate facts.
Who ever posted the story left out the "?" from the original submitter which changes the entire context of the article. The article itself is a speculation based on what Intel is rumored to want to do. There is not a confirmation that they are going to fab Apple's iOS chips.
Well, there's spam egg sausage and spam, that's not got much spam in it.
I strongly suspect that(even if PA semi weren't fabless) lack of backwards compatibility is seen as a feature, not a bug, by Apple. The last thing that they want is to make it easier for people to release warmed-over desktop applications for their precious touch-based platforms. Trying to make a handheld desktop is more or less what made WinCE such a pain in the ass to use. Apple would likely deliberately break compatibility before putting up with that, even if it could be enabled without switching architectures.
that's 140 million dollars in sales in the most recent quarter on iPads alone (nevermind iphones or ipods). You may call that drop in the bucket if you want, but I don't Samsung is going to.
Sure, but they sold over 10 million of the Android phones Apple is pissed about (Galaxy S), at what, 400 bucks each? So that's over twice as much, for one model alone.
When Samsung sells a part for $X to a supplier, they get $X. When a Samsung phone is sold for $Y, they do not collect $Y. The phone is sold through other entities, who also choose to profit.
Apple is reported to be buying 7.8 billion from Samsung this year alone through existing contracts. This makes them their largest customer. So yeah, it's a big deal. Probably worth settling lawsuits and licensing patents over. Time will tell.
Hint: Intel != x86
Apple buys chips for their iDevices on a seriously awesome scale, in advance. They have a epic boatload of cash which they drop on suppliers, and get very, very good deals as a result.
Intel has serious fab capacity, and although this may not really fit with their "x86" or "Atom" strategy, maybe they wouldn't mind having some more BILLIONS dumped on them to produce epic tons of chips for Apple?
Currently Samsung does a lot of that: Apple is their single biggest customer, but Samsung in another unit is directly competitng with Apple's biggest cash cows. This creates an interesting and problematic relationship. Although Intel has Atom strategies they want to get out there, and although Intel wants to move their chips into the mobile arena -- they aren't in the same business as Samsung, and aren't in the same business as Apple.
It doesn't *hurt* Intel's strategy to get this side-job and make a boat load of money, because this is a customer that they would never have had any chance at before. Its all billions of dollars gravy for them (if they have the fab capacity, which I wouldn't really doubt they do). And if in some future world they manage to make a super-cool, energy-efficient, fast little moblie processor that can really compete with the ARM cores -- they'll already have a good relationship with Apple to sell it to them.
That said: the site is slashdotted, so I have no idea if its real, and I'm doubtful until I see it from some more reputable sources. But, it isn't illogical to think it may be.
It'd be win-win for Intel.
Yup. The slashdot article and the winbeta article it references don't really unpack this, but this is about Intel's foundry business, not their x86 business. Presumably Intel will be making A5s in their foundry. It would be bizarre for Apple to switch away from the A4/A5 processor line after the investment they've made in it, particularly because there's simply no way the x86 architecture can ever go toe-to-toe with the ARM architecture on power efficiency.
simply no way the x86 architecture can ever go toe-to-toe with the ARM architecture on power efficiency.
Spoken like someone without a clue. There is fundamentally absolutely nothing in x86 that would cause it to consume more power than ARM. If anything the instruction predication in ARM gives x86 an advantage.
As ARM processors get more performance competitive with x86 they are beginning to match the power usage too. The big power advantage in current high performance ARM's is more due to the SOC integration than the architecture. Just wait, I will bet that in another generation or two, the roles will reverse as intel brings a much better fab process/integration, and the huge force of making a limited number of CPU models to bear against the dozens of ARM vendors each trying to optimize their particular design against a generic fab process. Samsung and Renesas might be the only ARM vendors with a chance, but even Renesas seems to prefer the SuperH.
Intel spends massive amounts on fab R&D and as a result are usually a node (generation) ahead of everyone else. Intel has had 32nm online and working for quite some time now. All Sandy Bridge chips are 32nm, many gen 1 Core i series laptops are 32nm, and so on.
Other fabs are catching up, GF will probably have 32nm chips coming out fairly soon for AMD, but Intel has been doing it for a long time, has scaled things up and has it working well. Also they are already building their 22nm fabs.
Only time Intel got outdone to an extent was with some companies doing a 40nm half-node. TSMC scaled down the 45nm process to 40nm and it is what all the GPU makers use now. Fine but it was fraught with problems and took a long time to get it working right and producing in volume. By that time Intel had 32nm parts on the market.
Same thing may happen again, a number of companies like TSMC are looking at skipping 32nm and going for a 28nm half node, based on 32nm scaled down. If they get that producing this year as they think they can, then they'll temporarily be ahead of Intel until Intel brings 22nm online.
However over all, Intel is always ahead on this shit. They spend a lot of money to stay that way.
I'm feeling particularly bored today..
BTW: Its not byte aligned instructions in x86 that cause the problems, but variable width instructions. Which of course ARM now supports (although not as bad as x86) via THUMB2, as well as the fact that the most recent ARM versions are also modal decoders, meaning that you have to know what mode the CPU is in before decoding a block of code.
Also, ARM has support byte load for as long as I can remember (always?), and added misaligned loads in ARM6 (IIRC). Its the misaligned load/store that generally causes pipeline problems not a partial word load/store. Furthermore, a really nasty thing that ARM has is load multiple, which pretty much can only be implemented efficiently (think exceptions during load/store) via microcode (or similar functionality).
BTW: Of the processors I write assembly on, ARM is probably my favorite. That said, its one thing to make a CPU consume a few tenths of a watt, its quite another to get what might be considered good performance at the same time. I get excited every time a higher performance ARM is announced, but I'm not sure people understand just how slow they really are. I wish the ARM vendors would start publishing SPEC CINT2006 numbers.
In general, the larger the die and transistor count, the higher the power requirements (not always true, but generally).
You of course remember that the 386 (fundamentally the same functionality provided by the base ARM instruction set) was implemented in 275 thousand transistors, and that the intel atom has roughly the same transistor count as the P4, yet burns significantly less power. Why is that? Well the first chapter of H&P talks about dynamic power (CMOS mostly burns power switching) being=.6CV^2f. Which initially looks like frequency is linear to power, but its more complex than that because as you reduce frequency you can reduce the V, which is squared! H&P then show that given a situation where you reduce the freq by 15% the power goes down by 60%. So bigger dies don't really mean anything when it comes to power, its more about frequency.
Looking at the current released data about the up and coming atoms (http://techreport.com/articles.x/18866/4) Intel is claiming battery life better than "leading smart phones" by about 10x.