iPhone 6s A9's and battery life

[Trader05]

http://www.macrumors.com/2015/10/07/tsmc-samsung-a9-battery-tests/

So in short, Apple has a 60/40 split between the A9 sourcing. Samsung A9's have less battery life for almost 2 hours on benchmark (or less optimized in the current iOS version i'm assuming) verses TSMC's.

You can find out what chip you have through this Lirum's Device Info Lite in the app store for free. *Apple removed from app store*

You can try this method posted by RichieRich1212 here:
http://www.iphonehacks.com/2015/09/how-to-identify-if-iphone-6s-has-a-tsmc-or-samsung-a9-chip.html

The reason I think this is crazy is how they just rush things out the door rather than testing and optimizing it first.

 

[kpkp]

Has anybody compared sustainable performance?

 

[richierich1212]

man Apple took notice and took down the app before I could download it!

 

[richierich1212]

Found another way to detect which version you have:

http://www.iphonehacks.com/2015/09/how-to-identify-if-iphone-6s-has-a-tsmc-or-samsung-a9-chip.html

I have a Samsung chip in my TMo 6s+ 64GB Space Grey...:|

 

[zaza]

man Apple took notice and took down the app before I could download it!

That's some shady stuff

I also imagine the TSMCs also produce less heat than the samsungs since lower battery usage means less energy is used which means less heat emitted, which in turn means less thermal throttling, which means TSMCs have more stable and better gaming performance than samsung chips.
Samsung always produces garbage, so this should be expected, their exynos chips are overheating underperforming garbage that only ignorant third world people would buy.

 

[kpkp]

Samsung always produces garbage, so this should be expected.

It's a bit early to blame everything on Samsung process, Apple (paper to metal) SoC team had to work on 2 processes, with 2 fabs meaning they had less resources which could lead to worse results (on one)?

 

[teejee]

It's a bit early to blame everything on Samsung process, Apple (paper to metal) SoC team had to work on 2 processes, with 2 fabs meaning they had less resources which could lead to worse results (on one)?

I agree, could just as well be Apples fault. And we might get a new stepping later that makes the Samsung SOC just as good as TSMC.

 

[QueBert]

Will suck if this pans out to be legit, I really want an 6s+ but if it's a crap shoot due to which chip it has, and there's no way to know. I'll stay away until Samsung or Apple, or both get their s**t together. I see this pushing up prices on Ebay and CL for verified non Samsung phones.

 

[Red Storm]

Will suck if this pans out to be legit, I really want an 6s+ but if it's a crap shoot due to which chip it has, and there's no way to know. I'll stay away until Samsung or Apple, or both get their s**t together. I see this pushing up prices on Ebay and CL for verified non Samsung phones.

Or pushing down prices for ones that have the Samsung SoC.

 

[mnewsham]

I agree, could just as well be Apples fault. And we might get a new stepping later that makes the Samsung SOC just as good as TSMC.

at the end of the day it IS apples fault, they elected to ship the product with samsung fabbed chips. Regardless of if samsung is or isn't crap compared to TSMC isn't really the point, apple made the final call and it's their product.

 

[poofyhairguy]

man Apple took notice and took down the app before I could download it!

That is as close as Apple comes to admitting fault in this kind of situation.

 

[Aikouka]

Will suck if this pans out to be legit, I really want an 6s+ but if it's a crap shoot due to which chip it has, and there's no way to know. I'll stay away until Samsung or Apple, or both get their s**t together. I see this pushing up prices on Ebay and CL for verified non Samsung phones.

The charts that I saw showed that a 128GB 6S+ had a very high chance of being a TSMC chip. (It may not be accurate, but the others seem to be close enough to 50/50 that it could be correct?)

 

[Artdeco]

Chipgate!!!!!!!

I'm curious if the advertised battery life was based on the SS or the TSMC chip. Have to look it up.

This would be an awesome thing for Anandtech to do a little research and testing on, like they did with the antenna issue.

http://youtu.be/pXmIQJMDv68

I have the SS chip in mine

 

[Commodus]

Sounds like it's a lot of sound and fury signifying nothing. You notice the difference in benchmarks, not in day-to-day use.

 

[Ns1]

if samsung battery life = quoted battery life then no issue.

 

[poofyhairguy]

This would be an awesome thing for Anandtech to do a little research and testing on, like they did with the antenna issue.

Apple already solved that problem via its hiring practices.

 

[Mopetar]

Apple is claiming a real-world 2% - 3% difference.

Considering that the whole 2 hour thing came from one person running one test on only two devices, it's been massively blown out of proportion, unless the only thing you do with your iPhone is run GeekBench all day.

 

[darkswordsman17]

Apple is claiming a real-world 2% - 3% difference.

Considering that the whole 2 hour thing came from one person running one test on only two devices, it's been massively blown out of proportion, unless the only thing you do with your iPhone is run GeekBench all day.

Well here's the thing, you say its getting blown out of proportion, but what data do you have to actually back that up? You can't say "yeah it's just one instance" and then provide nothing to dispute it.

Others kinda did. There's still a tangible difference though, and there's plenty of times where an extra 5-10% at the end of the day meant the difference between a dead battery and me still being able to make a call. Plus it also highlights Apple's lack of fast charging.

Oh, and plus those tests disputing there's a big difference didn't run most of the way down. How do we know that Apple doesn't have a software tick to try and even things out? Of course it could be the opposite and that over a full rundown we'd actually see them even out (or maybe even the Samsung hold out longer)?

We don't have enough real data. And yes, Apple is getting extra focus from this, but hey they want Apple bias, well they'll get it in return as well. Fact is, when people make a fuss about Apple screwing something up, it generally leads to improvements across the industry, so even if you have no interest in Apple, it could benefit you otherwise.

 

[mmntech]

Apple is claiming a real-world 2% - 3% difference.

Considering that the whole 2 hour thing came from one person running one test on only two devices, it's been massively blown out of proportion, unless the only thing you do with your iPhone is run GeekBench all day.

From user accounts, it's actually closer to 6-11%.
http://www.macrumors.com/2015/10/08/samsung-tsmc-a9-videos/

But yes, it has been overblown. It probably equates to a half hour less battery life between charges. That's not really going to impact your average user.

Curiously, GPU performance between is identical. The Samsung chips also use a smaller process: 14nm vs 16nm. Logically, you'd think the die shrink would improve efficiency.
http://www.anandtech.com/show/9665/apples-a9-soc-is-dual-sourced-from-samsung-tsmc

I wonder if they're dual sourcing the A9X as well. Be interesting to see how that goes. Guess we'll have to wait until next month to find out.

 

[mnewsham]

14nm vs 16nm

The name of their process node doesn't really directly relate in that way.

CPP :
Intel's 22nm Trigate is 90nm
intel's 14nm Trigate is 70nm
Samsung's 14nm FinFET is 78nm
TSMC's 16nm FinFET is 90nm

Metal pitch
intel's 22nm Trigate is 80nm
intels 14nm Trigate is 52 for their mobile line, 60nm for normal power chips
Samsung's 14nm FinFET is 64nm
TSMC's 16nm FinFET is 64nm

High Density SRAM cell area
Intel's 22nm Trigate is .1080um2
intel's 14nm Trigate is .0588um2
Samsung's 14nm FinFET is .0645um2
TSMC's 16nm FinFET is .0700µm2

Fin pitch
Intel's 22nm Trigate is 60nm
intel's 14nm Trigate is 42nm
Samsung's 14nm FinFET is 48nm
TSMC's 16nm FinFET is 48nm

 

[catinthefurnace]

The Samsung chips also use a smaller process: 14nm vs 16nm. Logically, you'd think the die shrink would improve efficiency.

Knowing the marketing name of the process is not enough information to determine much of anything.

 

[RaulF]

http://www.macrumors.com/2015/10/07/tsmc-samsung-a9-battery-tests/

So in short, Apple has a 60/40 split between the A9 sourcing. Samsung A9's have less battery life for almost 2 hours on benchmark (or less optimized in the current iOS version i'm assuming) verses TSMC's.

You can find out what chip you have through this Lirum's Device Info Lite in the app store for free. *Apple removed from app store*

You can try this method posted by RichieRich1212 here:
http://www.iphonehacks.com/2015/09/how-to-identify-if-iphone-6s-has-a-tsmc-or-samsung-a9-chip.html

The reason I think this is crazy is how they just rush things out the door rather than testing and optimizing it first.

Found another way to detect which version you have:

http://www.iphonehacks.com/2015/09/how-to-identify-if-iphone-6s-has-a-tsmc-or-samsung-a9-chip.html

I have a Samsung chip in my TMo 6s+ 64GB Space Grey...:|

Yeah, i'll right on that profile install on my phone...

 

[mmntech]

Knowing the marketing name of the process is not enough information to determine much of anything.

Care to explain for the laypeople? I was always told smaller transistors improved efficiency. A difference of 2nm is insignificant, but you'd think the smaller die would be the more efficient one.

 

[StrangerGuy]

Care to explain for the laypeople? I was always told smaller transistors improved efficiency. A difference of 2nm is insignificant, but you'd think the smaller die would be the more efficient one.

If you are talking about a decade ago, yes. Now transistors are so small that the choices of material, tuning the process to the chip design and vice-versa matters much more, with that said in recent times the actual feature sizes of the manufacturing process aren't even consistent at all with the so called xx-nanometers the fab companies market them as.

 

[Mopetar]

Well here's the thing, you say its getting blown out of proportion, but what data do you have to actually back that up? You can't say "yeah it's just one instance" and then provide nothing to dispute it.

I linked to an article that sources Apple as saying that the real-world performance difference comes down to 2-3%. Them being the ones capable of collecting user submitted data by the thousands-millions would probably put them in the best position to know that kind of thing.

I'll take that over one individual's test using two devices. For all you know he could have had the worst possible Samsung SoC and the best possible TSMC SoC, but because there were only one of each, you'd never be able to tell what kind of distribution they have in terms of performance or whether your results were even close to typical.

Curiously, GPU performance between is identical. The Samsung chips also use a smaller process: 14nm vs 16nm. Logically, you'd think the die shrink would improve efficiency.

Other threads have suggested Samsung is having yield problems. That would suggest that even though their chips meet Apple's expectations, TSMC is probably getting a lot of chips that could have been clocked higher and stayed within the power budget, but Apple doesn't go the usual Intel/AMD route where they bin chips into performance categories so you're a lot more likely to see greater variation in performance even across the same manufacturer.