But WHICH iPhone 6s do I have ?

[Couperin47]

So it turns out Apple has dual sourced the production of the A9 cpu used in the 6s between TSMC and Samsung, and because the processes these two use are different, so are the resulting chips, with considerable differences in performance and battery life. There is an app to inform you who's chip is inside and, basically, if you have a Samsung chip... you have been screwed: it's slightly slower, runs hotter and, of course, gives you easily 20% less battery life.
 
Details: http://apple.slashdot.org/story/15/10/08/1446218/not-all-iphone-6s-processors-are-created-equal?utm_source=rss1.0mainlinkanon&utm_medium=feed

 

[Dollar]

I've been pretty happy with my 6s, so I'm hoping I have the TSMC chip.  But I'm not going to check to see which one it is, because then there's a 50% chance I'll be upset with no way to change anything about it.

 

[swiftaw]

Use a free of charge app called Battery Memory System Status Monitor, which does exactly what the name says.

How to tell if your iPhone 6s has TSMC or Samsung chip
Step 1: Launch Battery Memory System Status Monitor on your device.

Step 2: Tap on the button labeled System at the top of the screen.

Step 3: Check your device’s ID number under the Device → Model sub-section:

N71mAP — iPhone 6s with TSMC-manufactured A9 processor
N66mAP — iPhone 6s Plus with TSMC-manufactured A9 processor
N71AP — iPhone 6s with Samsung-manufactured A9 processor
N66AP — iPhone 6s Plus with Samsung-manufactured A9 processor

 

[Senator Donut]

Oh noes

 

[DJnVa]

Use a free of charge app called Battery Memory System Status Monitor, which does exactly what the name says.

How to tell if your iPhone 6s has TSMC or Samsung chip
Step 1: Launch Battery Memory System Status Monitor on your device.

Step 2: Tap on the button labeled System at the top of the screen.

Step 3: Check your device’s ID number under the Device → Model sub-section:

N71mAP — iPhone 6s with TSMC-manufactured A9 processor
N66mAP — iPhone 6s Plus with TSMC-manufactured A9 processor
N71AP — iPhone 6s with Samsung-manufactured A9 processor
N66AP — iPhone 6s Plus with Samsung-manufactured A9 processor

 

None of those.
 
Mine is N56AP

 

[Oil Can Dan]

Is there any way to tell what you're going to get in advance of buying one?  My wife is about to pull the trigger upgrading from her 5 and obviously would prefer the better chip.  

 

[Couperin47]

Is there any way to tell what you're going to get in advance of buying one?  My wife is about to pull the trigger upgrading from her 5 and obviously would prefer the better chip.  

 
of course not, they are contemptuous of their customers, not that stupid

 

[teddykgb]

Try not to believe the Couperin anti apple propaganda machine.  Apple has been dual sourcing components for quite some time and Apple is claiming a variance of 2-3%.
 
http://techcrunch.com/2015/10/08/apple-says-battery-performance-of-new-iphones-a9-chips-vary-only-2-3/
 
The relevant parts:
 

he 2-3% difference Apple is saying it sees between the battery life of the two processors is well within its manufacturing tolerances for any device, even two iPhones with the same exact processor. In other words, your iPhone and someone else’s iPhone with the same guts likely vary as much as 3%, regardless of who made them.
Basically, if you can tell the difference in real-world usage between the two processors, you should take a Voigt-Kampff test.

 
There's really no reason to panic or worry about this nor is there really anything untoward going on here.

 

[wibi]

 
None of those.
 
Mine is N56AP

 
This about the 6s and that chip is a 6plus chip

 

[NortheasternPJ]

Try not to believe the Couperin anti apple propaganda machine.  Apple has been dual sourcing components for quite some time and Apple is claiming a variance of 2-3%.
 
http://techcrunch.com/2015/10/08/apple-says-battery-performance-of-new-iphones-a9-chips-vary-only-2-3/
 
The relevant parts:
 
 
There's really no reason to panic or worry about this nor is there really anything untoward going on here.

Isnt ChipGate more fun though????

 

[Couperin47]

Try not to believe the Couperin anti apple propaganda machine.  Apple has been dual sourcing components for quite some time and Apple is claiming a variance of 2-3%.
 
http://techcrunch.com/2015/10/08/apple-says-battery-performance-of-new-iphones-a9-chips-vary-only-2-3/
 
The relevant parts:
 
 
 
 
There's really no reason to panic or worry about this nor is there really anything untoward going on here.

 
My original post is 100% factual. You. otoh only cite Apple's response which is, so far, not borne out by anyone independently, which simply makes you an Apple promotional sock puppet.
 
Dual sourcing of any cpu used to be the norm...manufacturers never wanted to be at the mercy of a single supplier. This attitude has disappeared as it became a practical impossibility.  Generally dual source cpus were identical in all meaningful respects. In this case Samsung is a half step behind the curve and most of these results were predictable, Apple specs the phone for the slower cpu, as others have in the past, claim everyone got what was promised and half of you are bitching because you get a product that performs even better? Presented as information and so that consumers can try and find a strategy to get the best value for money
 
Dell and others commonly dual source the led panels used in their computer monitors and the panels from different makers are usually quite different. Much bigger issue.

 

[teddykgb]

Couperin, my post was intended as a bit of a light hearted ribbing as you're clearly one of those old school windows guys but rereading it it definitely doesn't come across that way, my apologies.
 
I don't think there's anything "significant" that's been proven to be an issue.  There's not much scientific about people running an app and benchmarks -- we're getting anecdotal data here.  Apple's response is pertinent because the assertion is that you might see a bigger gap between two TSMC chips than you would between a TSMC and a Samsung Chip.  These types of manufacturing tolerances are common within the industry and I think people going nuts over this without real proof is just another attempt at having an apple related scandal to drive advertising.

 

[Byrdbrain]

The article linked in the OP actually shows that the Samsung chip is more advanced than the TSMC. The Samsung chip has a smaller die size(96 mm2 vs 104mm2) due to the 14 nm line width vs the 16nm line width of the TSMC chip.
The performance of the Samsung chip may not be as good(though it should be better) but it isn't because it was fabbed with inferior technology. What almost certainly happened is Samsung couldn't supply all the chips needed so Apple had to second source from TSMC.
Benchmarking Point of View
 
Edit: From the article linked-
From a benchmarking point of view, the smaller die size shows a leadership in technology scaling for Samsung. 

 

[derekson]

Die size isn't everything. A better implemented 16 nm process could outperform an inferior 14 nm process. And that appears to be the actual case that we're seeing here. In fact when the news originally broke of the dual-sourced A9, people assumed the Samsung chip would be the superior option due to the better process on paper.
 
Ultimately there is significant variation in all chips due to the nature of production of the wafers and needing to have a reasonable yield. You can't reject every chip that doesn't come out with the best leakage and voltage. The real questions are ultimately: 
A. Is the variance between the average TSMC and Samsung chip really significantly bigger than the expected variance we'd see between better and worse examples of a single sourced chip? It's also possible that TSMC's process simply has a tighter variance and that good Samsung chips are just as good (or better than, owing to the 14 nm process) as TSMC chips. 
B. Are Apple's battery life claims based on a worst case scenario (e.g. a Samsung A9 that was on the lower end of passing QC)
C. How much does variance in the chip affect battery life versus variance in other key components like the LED backlight and the battery itself? It's quite possible that even with a less efficient example of the A9 you've still got good battery life since you "won the battery lottery" and or the "backlight lottery". 
 
FWIW my IPhone 6S Plus has a Samsung chip according to the app script some dude posted to test it. But honestly my battery life has been superlative, so I likely got a better example of the battery and/or the backlight or whatever other components significantly affect battery life. Or maybe I just got one of the better examples of a Samsung chip. There are a TON of variables at play here; a modern smartphone is a complex device with a lot of components and basically every single component has significant variance due to manufacturing processes.
 
I also suspect 9.1 improves battery life a good bit (I've been running the beta from day 1 since my backup of my old phone was on 9.1 beta and i needed to upgrade to restore from that backup). 9.1 seems to be better optimized than 9.0, at least as far as performance goes, so I'd guess it probably does better for battery life as well. 

 

[derekson]

Anandtech has a decent piece up with a rational look at this situation:
 
http://anandtech.com/show/9708/analyzing-apple-statement-for-tsmc-and-samsung-a9

 

[derekson]

 
 
 
Dell and others commonly dual source the led panels used in their computer monitors and the panels from different makers are usually quite different. Much bigger issue.

 
There was a big hullaballoo about the dual sourced LCD panels in the first renditions of the retina MacBook Pro. The LG panels were apparently pretty crummy, with image retention and ghosting in a lot of examples after a few months of use.
 
In the 2014 rMBP models (which I own an example of) there was also a situation where the two different SSD suppliers had significantly different performance on benchmarks. The Samsung SSD was something like 200 MB/s faster than the Sandisk, if I remember correctly.

 

[Murderer's Crow]

This is just an issue that requires a bit of common sense and taking a step back because nothing here is scientific or difficult to understand with benchmarking and production issues. If there was a 20% variance, people would be going nuts about how poor their battery life is. Period.

Tired of every single apple release becoming an attempt for anyone and everyone to turn it into a scandal.

 

[Byrdbrain]

Die size isn't everything. A better implemented 16 nm process could outperform an inferior 14 nm process. And that appears to be the actual case that we're seeing here. In fact when the news originally broke of the dual-sourced A9, people assumed the Samsung chip would be the superior option due to the better process on paper.
 
Ultimately there is significant variation in all chips due to the nature of production of the wafers and needing to have a reasonable yield. You can't reject every chip that doesn't come out with the best leakage and voltage. The real questions are ultimately: 
A. Is the variance between the average TSMC and Samsung chip really significantly bigger than the expected variance we'd see between better and worse examples of a single sourced chip? It's also possible that TSMC's process simply has a tighter variance and that good Samsung chips are just as good (or better than, owing to the 14 nm process) as TSMC chips. 
B. Are Apple's battery life claims based on a worst case scenario (e.g. a Samsung A9 that was on the lower end of passing QC)
C. How much does variance in the chip affect battery life versus variance in other key components like the LED backlight and the battery itself? It's quite possible that even with a less efficient example of the A9 you've still got good battery life since you "won the battery lottery" and or the "backlight lottery". 
 
FWIW my IPhone 6S Plus has a Samsung chip according to the app script some dude posted to test it. But honestly my battery life has been superlative, so I likely got a better example of the battery and/or the backlight or whatever other components significantly affect battery life. Or maybe I just got one of the better examples of a Samsung chip. There are a TON of variables at play here; a modern smartphone is a complex device with a lot of components and basically every single component has significant variance due to manufacturing processes.
 
I also suspect 9.1 improves battery life a good bit (I've been running the beta from day 1 since my backup of my old phone was on 9.1 beta and i needed to upgrade to restore from that backup). 9.1 seems to be better optimized than 9.0, at least as far as performance goes, so I'd guess it probably does better for battery life as well. 

I agree that die size isn't everything and most everything else you say as well I was mainly pointing out that the link in the OP doesn't support the OP other than to say there are two suppliers.
I haven't done any other research into this issue other than clicking on that link as I don't have a 6S and don't plan on getting one.

 

[derekson]

This is just an issue that requires a bit of common sense and taking a step back because nothing here is scientific or difficult to understand with benchmarking and production issues. If there was a 20% variance, people would be going nuts about how poor their battery life is. Period.

Tired of every single apple release becoming an attempt for anyone and everyone to turn it into a scandal.

 
I happen to find it a fairly interesting issue as someone into tech stuff, but yeah the OP's concern trolling certainly blows any actual issue out of proportion. In particular the "easily 20% less battery life" claim seems a bit inflated, to say the least.
 
In the end it's pretty much the same as the "Bendgate" thing from last year: something that people found and wrote clickbait articles about to try and stir up a controversy when it's really a pretty insignificant concern and something all phones have to some extent.

 

[derekson]

So Tom's Hardware did a test, and their Samsung chip was actually more efficient:
 
http://www.tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html
 
I'd say the most likely hypothesis at the moment is that the Samsung chip is probably just higher variance than the TSMC.

 

[Couperin47]

So Tom's Hardware did a test, and their Samsung chip was actually more efficient:
 
http://www.tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html
 
I'd say the most likely hypothesis at the moment is that the Samsung chip is probably just higher variance than the TSMC.

 
LOL you can't have a smaller sample size (they tested exactly one of each phone) and while the overall performance was so small it clearly falls within variations you will probably see between any 2 phones with the same processor...the, admittedly simplistic,  battery life test showed a variation of over 10%. What does that prove ? I have no idea, and neither do you as, for example, we have no idea whatever as to the variation in the batteries themselves.

 

[derekson]

Consumer reports also finds no measurable difference between Samsung and TSMC chips in iPhone 6S models.
 
Looks like there's something particular about the GeekBench 3 battery test that is throwing the Samsung A9 phones for a loop. No idea what that might be. But it seems to be the only test generating a significant difference, between real world usage, other benchmarks, and simulated real world usage tests.

 

[Murderer's Crow]

You have to break that news to Couperin more slowly. Too many emotions too fast.

 

[amlothi]

https://youtu.be/m1fUil7QLNI

 

[Couperin47]

You have to break that news to Couperin more slowly. Too many emotions too fast.

 
You have a vivid and ridiculous imagination if you think I get 'emotional' about anything Apple churns out. Stop projecting.

 

[derekson]

https://youtu.be/m1fUil7QLNI

 
I stopped watching that as soon as he compared processor die size to the size of the sensor on a digital camera and claimed that a larger die chip normally means better performance and efficiency. I was worried my palm would hit my face hard enough to give me a concussion if I continued.

 

[Apisith]

Dude is losing his fastball. His videos from a year or so ago were better sourced and had stuff that you wouldn't find elsewhere.

 

[crystalline]

So Tom's Hardware did a test, and their Samsung chip was actually more efficient:
 
http://www.tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html
 
I'd say the most likely hypothesis at the moment is that the Samsung chip is probably just higher variance than the TSMC.

How does fab variance lead to different battery life and performance? I can see performance if they clock a chip down to deal with poor quality traces/features. Is battery life degradation a consequence of extra capacitance leading to running at higher voltage?

 

[derekson]

How does fab variance lead to different battery life and performance? I can see performance if they clock a chip down to deal with poor quality traces/features. Is battery life degradation a consequence of extra capacitance leading to running at higher voltage?

 
My understanding is that it's mainly a consequence of gate leakage due to imperfections in the dielectrics. I'm hardly an expert though, and it's hardly a simple topic.

 

[mt8thsw9th]

You have a vivid and ridiculous imagination if you think I get 'emotional' about anything Apple churns out. Stop projecting.

Yeah, you've been largely ambivalent from what I've read.

 

[Papelbon's Poutine]

Yeah, you've been largely ambivalent from what I've read.

I'm pretty sure he owns stock in Apple and is just trying to drive it down to buy more. Then he'll get on board. Otherwise he's totally middle of the road.

 

[Couperin47]

I'm pretty sure he owns stock in Apple and is just trying to drive it down to buy more. Then he'll get on board. Otherwise he's totally middle of the road.

 
My last investment in a tech stock was in a company who's products I used at the time: Syquest. I doubt anyone here knows or remembers what they made and in the end they were a total bust in 1998.

 

[Byrdbrain]

 
I stopped watching that as soon as he compared processor die size to the size of the sensor on a digital camera and claimed that a larger die chip normally means better performance and efficiency. I was worried my palm would hit my face hard enough to give me a concussion if I continued.

Yeah that is as far as I made it in the video, as soon as he said that I knew he had no idea what he was talking about.

 

[cgori]

 
My last investment in a tech stock was in a company who's products I used at the time: Syquest. I doubt anyone here knows or remembers what they made and in the end they were a total bust in 1998.

Last summer i rebuilt an Linux box (dual PPro-200, w00t) to recover 15 different syquest 270mb carts full of crap that i had sitting in storage. I had a hell of a time cobbling back together the correct scsi termination, but all save 1 cartridge still worked.

Iomega ate them for breakfast

 

[Doc Zero]

I test and review smartphones for a living and while I'm—cough—not ready to say that I agree with everything in that Chipgate video posted upthread, I can say that the overall conclusion they're arriving at pretty much reflects what we've seen in our testing.

 

[Couperin47]

Last summer i rebuilt an Linux box (dual PPro-200, w00t) to recover 15 different syquest 270mb carts full of crap that i had sitting in storage. I had a hell of a time cobbling back together the correct scsi termination, but all save 1 cartridge still worked.

Iomega ate them for breakfast

 
Irony: after the bankruptcy...guess who bought the remains of Syquest ? Yeah...Iomega. Of course once the Zip became obsolete too, they faded into irrelevancy also.

 

[The_Powa_of_Seiji_Ozawa]

I'm pretty sure he owns stock in Apple and is just trying to drive it down to buy more. Then he'll get on board. Otherwise he's totally middle of the road.

 
He found a coupon for 10% off Apple stock and free shipping

 

[The_Powa_of_Seiji_Ozawa]

 
Irony: after the bankruptcy...guess who bought the remains of Syquest ? Yeah...Iomega. Of course once the Zip became obsolete too, they faded into irrelevancy also.

 
Got them in storage next to the CD-WORM drive and the 2.88mb super floppy drive.

 

[Couperin47]

 
Got them in storage next to the CD-WORM drive and the 2.88mb super floppy drive.

 
God how we used to fight for something resembling fast storage... how many tried to low level reformat Seagate 225 doorstops to RLL to eek 30 mb out of a 20 mb drive, and with the inferior platter coatings it never really worked....

 

[cgori]

Definitely remember re-interleaving a 10mb full height drive from 6:1 to 5:1 to squeeze out a tiny bit of performance.

I never had one of those RLL controllers but always wanted one (Perstor, I think?)