iPhone 6s Uses Different-Sized A9 Chips From Samsung and TSMC

[perkedel]

So how do we find out which one is which? For curiousity and educational purposes.
There are those who say their phone is running hot, is that the tsmc version?

 

[Norbs12]

So how do we find out which one is which? For curiousity and educational purposes.
There are those who say their phone is running hot, is that the tsmc version?

EDIT: misread that, no way outside of opening it as far as we know for now. As far as the phones running hot, I doubt t has to do with the die size more likely a software issue.

 

[wicker21]

I don't care which is in my phone as long as it keeps working as fast as it does. Kicks the crap out of the previous generation and any current Android.

 

[btbeme]

I'll take a TSMC chip, thank you. Samsung can go whiz up a rope for all I care. No Samsung anything in my house if I have any choice.

 

[Simoleon547]

It's not a big deal. But if I could get an extra 20 minutes of battery life every day for 2 or 3 years, I'd be sorely tempted to find an excuse for an exchange.
Still, people would have to find a way to detect their chip, maybe by TDP testing, throttling tests, low level voltage info, chip identifier tags... And there isn't an app for that. Yet.

 

[uid15]

So how do we find out which one is which? For curiousity and educational purposes.
There are those who say their phone is running hot, is that the tsmc version?

Take it apart, remove the solder paste and desolder it with IR reflow; all the best.

PS, if you find one of these inside, you will know you've been had:

 

[Simoleon547]

Why is this a story? Different processes, different foundries, different sizes. It's be more newsworthy if the were exactly the same size.

Because one is going to have slightly less battery life to get the same performance.

 

[cdm283813]

If Apple can lie on stage about the iPad mini 4 being just as powerful as the Air 2 then I can definitely see them pulling this crap. I wonder if the "early" reviewers got the cherry picked Samsung made chips.

 

[Simoleon547]

"Introducing, the iPhone 6S*"

 

[uid15]

Because one is going to have slightly less battery life to get the same performance.

Battery usage is not a constant. It is not perfectly uniform across all the iPhones worldwide. Usage differs (sorry to insult your intelligence by being a typical "spell-out-the-blindingly-obvious" dork).

 

[cdm283813]

I'll take a TSMC chip, thank you. Samsung can go whiz up a rope for all I care. No Samsung anything in my house if I have any choice.

What about the other chips used in your iPhone? Should just toss your 6S and other Apple gear out the windows then.

 

[mi7chy]

Totally explains why some are overheating doing almost nothing and others are fine. How do you pick out the better 14nm one?

 

[Rogifan]

So maybe I missed something. Where was it confirmed that the tsmc chip caused overheating and battery issues?

 

[skinned66]

 

[KPOM]

Because one is going to have slightly less battery life to get the same performance.

I highly doubt it will be noticeable. We could be talking about a couple of minutes. Tops.

 

[KPOM]

Totally explains why some are overheating doing almost nothing and others are fine. How do you pick out the better 14nm one?

Seriously? The iPhone 6 used an A8 built on a 21nm fab. The Apple Watch uses a 28nm process. Why aren't these catching fire?

 

[Simoleon547]

So maybe I missed something. Where was it confirmed that the tsmc chip caused overheating and battery issues?

It won't. The difference in heat at a certain performance will be minimal. The battery life won't be hurt too much, a big chunk of battery power goes to the screen and other chips on the mainboard. It's probably less than half that goes to the A9. If the processes had the same power draw per mm^2, then the A9 power draw would be higher by the same percentage increase in chip area, and consequently that % multiplied by the % of power that the A9 chip draws of the total power would equal the percent loss in battery life.

 

[nutjob]

Because one is going to have slightly less battery life to get the same performance.

Don't know if you're serious, but in actual fact every chip has a slightly performance profile. So some are going to be more efficient than others. Chips are "binned" according to their performance and Apple may actually let a lot of variation through, because it costs money to throw away chips and it doesn't care too much about how any particular phone performs.

 

[nutjob]

I highly doubt it will be noticeable. We could be talking about a couple of minutes. Tops.

You have no basis for saying that at all. There could be a wide variation in the chips made by the same foundry. It depends on what Apple specifies. Going by how cheap they are with RAM, I'd say there is a wide variation in power usage.

 

[flat five]

Totally explains why some are overheating doing almost nothing and others are fine.

yep.. totally!
fine piece of logic right there :thumbsup:

(i kid.. i kid)

 

[MJD Dawg]

DIFFERENT-SIZED-CHIP-GATE!!!

I'm just going with Chip Gate.

Damn you Apple.

 

[macfoxpro]

10 years of VLSI semiconductor design and fabrication, primarily II-VI and III-V but also with some Group 4 experience. I have a master's degree from Stanford in Materials Science and took graduate level courses in semiconductor physics and semiconductor device fabrication. I also have a master's degree in Optical Physics from the University of Arizona.

The size of the die is irrelevant. The mask set (i.e. the schematic of the chip itself) is almost certainly identical between the two suppliers.

It is possible that one chip could run hotter than the other chip. It has to do with the lateral doping profile and transistor geometry and a bunch of other things which probably are different between the suppliers. Of course that's also true for chips from the same supplier. How much different, no one but Apple could know. It could be that they're so similar, the performance is identical. Without a direct statement from Apple, it will forever be idle speculation.

Dang. This guy!! So much for CNTRL>ALT>DELETE

By the way, what's the square root of the Apple Campus?
jk....
Thanks for the info... whatever it means.

 

[Jimmy James]

Chip-Gate. I'm sure I'm the first person to think of this. Right?

 

[uid15]

I'm glad someone emailed all the silicon engineers and invited them to comment on the thread, otherwise this whole topic would be a farce.

</sarc>

 

[themyst]

Wasn't there a way to identify the hardware SKU of NAND to determine if it was TLC/MLC back in the 6/6+ NAND crash days? Couldn't that same method be used to identify which A9 is in a specific iPhone?

The 14nm process will draw less power than the 16nm process. Fact.