Just to nitpick the wording in the article, is a ten percent size reduction really all that "significant"? i am assuming all the chip components are reduced proportionally.
Except that every single tear down done to date shows a TSMC processor. I'd say that rumor is debunked, unless you can find a tear down with a Samsung processor.
Surely with Samsung supplying 40% of the chips you'd think SOMEONE would have seen one by now.
On the other hand, Chipworks discovered TSMC and Samsung A9 processors right away.
Except that every single tear down done to date shows a TSMC processor. I'd say that rumor is debunked, unless you can find a tear down with a Samsung processor.
Surely with Samsung supplying 40% of the chips you'd think SOMEONE would have seen one by now.
On the other hand, Chipworks discovered TSMC and Samsung A9 processors right away.
Elsewhere it's been indicated that the samsung chip is in the 6S and the TSMC is in the 6S Plus.
There's not going to be enough of a difference. For the sake of argument I'm running geekbench's battery-life benchmark with all the radios on to see what the battery life ends up being on the 6S 128GB, but I'm willing to bet that nobody who actually uses the phone would notice which chip would be used.
So now are we going to get chipgate where people demand to know which chip is in their phone and if their phone runs a bit hot or they encounter battery life issues they'll be demanding Apple swap out their phone for one with the smaller chip. It wouldn't surprise me. :rolleyes:
Most Apple iPhone customers want to buy an iPhone.
I certainly don't know much about chip baking, but didn't expect this. I wonder if the larger one runs any cooler than the smaller one. Or if additional airspace around the smaller one aids in cooling. (???)
It works the other way round. Smaller process size means less energy required to do the work.
Quote:
Originally Posted by dasanman69
It would be quite ironic if people demanded the one with the Samsung chip.
We already had that when the original 15" MBPR was released. All the screens made by LG developed serious flaws and people were anxiously typing stuff in the terminal to find out whether they had a Samsung or LG screen. I was one of them. Luckily mine had a Samsung.
I don't think anybody knows. It is interesting that ChipWorks got two different chips so quickly into the product launch. You would think that locality of orders would result in batches of one vendors processor going through the lot.
I highly doubt that anybody at Apple looks at vendors like this. Rather the look for assurances that production demand can be meant which for Apple is increasingly a problem. This was a long time ago but at one time Apple was taking >80% of Samsungs capacity and shortly after that rumor a factory extension was done in Texas. Demand for Apples A series processors is massive now, there is a very good possibility that they need both partners just to have a bit of breathing room to get all the chips they need.
Also if you follow the chip industry news there was a report a year or two ago that Global Foundries plant was validated to build Apples chips due to the risk of running out of capacity.
There are strong rumors that TSMC will have 10 nm class chips in late 2016, which of course would be an even bigger up grade to the watch. Every micro watt makes a difference on the watch. The other thing is this: are yearly updates required or even desirable on a watch.
Long term, yearly updates make little sense, but short term, Apple needs to crush and assert their dominance on the competition; not leave them any chance. If that means better process, more sensors, GPS, more compact watch, better materials, more colors, whatever, they'll do it.
They didn't do that with the phones and that let Android grow into a monster. With the watch, I can see them improving rapidly over 3-4 generations and then maybe going for longer update cycles 2-3 years (a bid like the Ipad which took 6 iterations over 5 years to get to this point.
I certainly don't know much about chip baking, but didn't expect this. I wonder if the larger one runs any cooler than the smaller one. Or if additional airspace around the smaller one aids in cooling. (???)
It works the other way round. Smaller process size means less energy required to do the work.
Quote:
Originally Posted by dasanman69
It would be quite ironic if people demanded the one with the Samsung chip.
We already had that when the original 15" MBPR was released. All the screens made by LG developed serious flaws and people were anxiously typing stuff in the terminal to find out whether they had a Samsung or LG screen. I was one of them. Luckily mine had a Samsung.
I could be wrong but I'm quite sure the chip is not a smaller "process" size. Just simply the body of the soc has more mass. If it was as you suggest, then the two phones would be completely different from each other. (battery life, etc.)
Edit: Ok, I just re-read. They are suggesting one is 14nm the other is 16nm process. I dunno. Doesn't make sense to me. Unless maybe one is for the s and one is for s+ (I'm no engineer)
Comments
Hey, you seem like a different person in recent months. Bought AAPL?
Hey, you seem like a different person in recent months. Bought AAPL?
Gatorguy's big day is tomorrow...
He's a happy guy!
Here's a headline wrt A9 production;
https://www.macnn.com/articles/15/08/26/engineer.hired.by.korean.company.revealed.finfet.fabrication.process.130131/
Samsung just doesn't know any other way, it would seem.
i am assuming all the chip components are reduced proportionally.
Except that every single tear down done to date shows a TSMC processor. I'd say that rumor is debunked, unless you can find a tear down with a Samsung processor.
Surely with Samsung supplying 40% of the chips you'd think SOMEONE would have seen one by now.
On the other hand, Chipworks discovered TSMC and Samsung A9 processors right away.
There's not going to be enough of a difference. For the sake of argument I'm running geekbench's battery-life benchmark with all the radios on to see what the battery life ends up being on the 6S 128GB, but I'm willing to bet that nobody who actually uses the phone would notice which chip would be used.
Most Apple iPhone customers want to buy an iPhone.
Huh. That is surprising.
I certainly don't know much about chip baking, but didn't expect this. I wonder if the larger one runs any cooler than the smaller one. Or if additional airspace around the smaller one aids in cooling. (???)
It works the other way round. Smaller process size means less energy required to do the work.
It would be quite ironic if people demanded the one with the Samsung chip.
We already had that when the original 15" MBPR was released. All the screens made by LG developed serious flaws and people were anxiously typing stuff in the terminal to find out whether they had a Samsung or LG screen. I was one of them.
No, I cant remember who was it that swear their life Apple wont be dual sourcing.
Quote:
I posted some rough figure earlier, but this year Apple will likely ship more 14/16 node transistor then Intel on their 14nm node.
So this is huge volume.
I don't think anybody knows. It is interesting that ChipWorks got two different chips so quickly into the product launch. You would think that locality of orders would result in batches of one vendors processor going through the lot.
I highly doubt that anybody at Apple looks at vendors like this. Rather the look for assurances that production demand can be meant which for Apple is increasingly a problem. This was a long time ago but at one time Apple was taking >80% of Samsungs capacity and shortly after that rumor a factory extension was done in Texas. Demand for Apples A series processors is massive now, there is a very good possibility that they need both partners just to have a bit of breathing room to get all the chips they need.
Also if you follow the chip industry news there was a report a year or two ago that Global Foundries plant was validated to build Apples chips due to the risk of running out of capacity.
There are strong rumors that TSMC will have 10 nm class chips in late 2016, which of course would be an even bigger up grade to the watch. Every micro watt makes a difference on the watch. The other thing is this: are yearly updates required or even desirable on a watch.
Long term, yearly updates make little sense, but short term, Apple needs to crush and assert their dominance on the competition; not leave them any chance. If that means better process, more sensors, GPS, more compact watch, better materials, more colors, whatever, they'll do it.
They didn't do that with the phones and that let Android grow into a monster. With the watch, I can see them improving rapidly over 3-4 generations and then maybe going for longer update cycles 2-3 years (a bid like the Ipad which took 6 iterations over 5 years to get to this point.
[IMG]http://forums.appleinsider.com/content/type/61/id/63465/width/200/height/400[/IMG]
Chipworks stated they found the two processors in two identical phones.
But wouldn't that imply that the motherboards would have to be slightly different as well? Or is the pin configuration in the exact same positions?
If different, that seems to me to be highly inefficient across the board - different parts, different boards, different repairs, etc.
Huh. That is surprising.
I certainly don't know much about chip baking, but didn't expect this. I wonder if the larger one runs any cooler than the smaller one. Or if additional airspace around the smaller one aids in cooling. (???)
It works the other way round. Smaller process size means less energy required to do the work.
It would be quite ironic if people demanded the one with the Samsung chip.
We already had that when the original 15" MBPR was released. All the screens made by LG developed serious flaws and people were anxiously typing stuff in the terminal to find out whether they had a Samsung or LG screen. I was one of them.
I could be wrong but I'm quite sure the chip is not a smaller "process" size. Just simply the body of the soc has more mass. If it was as you suggest, then the two phones would be completely different from each other. (battery life, etc.)
Edit: Ok, I just re-read. They are suggesting one is 14nm the other is 16nm process. I dunno. Doesn't make sense to me. Unless maybe one is for the s and one is for s+ (I'm no engineer)