One is probably already running hotter than the other one!
http://www.technobuffalo.com/2015/09/27/iphone-6s-overheating-for-some-disabling-led-flash/
"FOR SOME"—i.e. for those with the "TSMC version" of the iPhone 6s/6s Plus?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iPhone 6s Uses Different-Sized A9 Chips From Samsung and TSMC
- Thread starter MacRumors
- Start date
- Sort by reaction score
http://www.technobuffalo.com/2015/09/27/iphone-6s-overheating-for-some-disabling-led-flash/
"FOR SOME"—i.e. for those with the "TSMC version" of the iPhone 6s/6s Plus?
Where's the video of two 6S+'s side by side showing how one is snappier than the other? lol Too much!
So... let me get this straight. A company sourced a component from two different suppliers? Noooooooo! /s.
Seriously though, most likely the 6s Plus uses the 16 nm TSMC chip. I ran several 3D Mark tests comparing my 6s Plus to an aggregate of other devices and there seems to be about a 1.2 percent difference in graphics. However, having a device that breaks 40k is AMAZING. I don't give a bleep about a POSSIBLE percent difference. It may turn out that over time as more data is collected it will be even narrower than 1 percent, and we should consider a 6s Plus has a higher pixel count to be managed. So a slight difference is normal. My battery life is near identical to my old 6 Plus. It may be a tad better, not sure. So, in my opinion, this means absolutely nothing.
Seriously though, most likely the 6s Plus uses the 16 nm TSMC chip. I ran several 3D Mark tests comparing my 6s Plus to an aggregate of other devices and there seems to be about a 1.2 percent difference in graphics. However, having a device that breaks 40k is AMAZING. I don't give a bleep about a POSSIBLE percent difference. It may turn out that over time as more data is collected it will be even narrower than 1 percent, and we should consider a 6s Plus has a higher pixel count to be managed. So a slight difference is normal. My battery life is near identical to my old 6 Plus. It may be a tad better, not sure. So, in my opinion, this means absolutely nothing.
lol processor manufacturers strive for smaller die sizes for many reasons. It can be looked up if you want to no one needs to demonstrate anything.
The smaller the die size, the less heat the chip produces, they less energy is needed to run it, the less heat a chip makes, the faster you can run it. Those are pretty important in a mobile device.
If it's "perfectly acceptable," then why didn't Tim Cook announce this when he announced the device at the keynote?
Why do we have to find out about it through the work of teardowns performed by Chipworks or others?
When you make millions of anything, there will be a tiny percent with problems. That same error happened before on a tiny percent of units. The report concluded that it doesn't affect many units. I'm not sure why you are trying to correlate that to a different SoC manufacturer.
I found this, according to iFixit the 6S plus would have the bigger die size (16nm) while the 6S has the smaller at least in their experiences.
"iFixit's teardowns found the Samsung version of the A9 in the iPhone 6S and TSMC's version in the iPhone 6S Plus, which makes sense—a larger phone has more room to spare for a larger chip—but that doesn't necessarily mean that all of the phones are being put together this way. In our testing, both the iPhone 6S and 6S Plus benchmarked nearly identically, and both behaved well during Geekbench's thermal throttling test."
http://arstechnica.com/apple/2015/0...ng-and-tsmc-are-making-the-a9-chip-for-apple/
Because they never talk about who makes their components. Maybe there are two different RAM manufacturers as well? Does it matter?
And are you going to submit a formal request to Apple to ask for an iPhone 6S full schematic, engineering CAD files, a list of all the individual suppliers of SMT resistors, caps, inductors, oscillators, semi, waveguides, delay lines etc... ALL OF WHICH ARE SUBSTITUTED WITH GREAT REGULARITY due to design tweaks, schematic revisions and versioning, supply logistics economics, assembley location etc, and NONE of which are ever brought to light, because it doesn't matter?
Sorry, end users, but you really are NOT entitled to be privy to this information, and even if you WERE, you wouldn't have the first idea of what to do with it.
Stick to buying and consuming, and leave the engineering to engineers
I sa
No not at all.
It more like a stepping. Though you buy a chip A, it may have stepping 99 or say 91. It could be a number of things. One being a shrink of the die, modifications to a particular layer or pinout.
One known benefit of the shrink is power consumption is usually improved. 16nm to 14nm is pretty significant.
Yes you can get the same performance specifications.
One will be more efficient.
Assuming the are using 300mm wafers the smaller die may yield 8-10% more die per wafer.
No not at all.
It more like a stepping. Though you buy a chip A, it may have stepping 99 or say 91. It could be a number of things. One being a shrink of the die, modifications to a particular layer or pinout.
One known benefit of the shrink is power consumption is usually improved. 16nm to 14nm is pretty significant.
Yes you can get the same performance specifications.
One will be more efficient.
Assuming the are using 300mm wafers the smaller die may yield 8-10% more die per wafer.
Last edited:
Not all all. Intel and AMD are two entirely different architectures.
Not entirely true. Intel and AMD both use the x64-64 architecture spec, which was invented by AMD.
BL.
Better not tell anyone here about how silicon parts are tested, binned, and features disabled at the silicon trace level and branded as another lesser IC because they didn't make the grade - shhh, there'll be an uproar,
LMAO!I probably won't lose sleep at night if I don't know who makes the CPU inside my iPhone… But, if I'm going to pay upwards of $700 for a device, don't I have the right to know who the maker of the major components (like the CPU) is? …And, especially if it doesn't matter, it shouldn't have to be kept a secret until Chipworks reveals all!
I'm not asking for the blueprints for the iPhone so I can build my own—I just feel I should be able to know which company made something as important as the CPU inside my phone.
Not AT ALL true, in fact. Same arch, diff manufacturers, slightly diff instructions sometimes and subsets of features.
X86 is x86, no matter whether it's Cyrix, VIA, Intel or AMD.
Yes, because one nanogram of glue and … THE freakin' CPU … both matter the same in a Smartphone… Smart comparison...
Yes, because one nanogram of glue and … THE freakin' CPU … both matter the same in a Smartphone… Smart comparison...
In fact, based on the part code, TSMC appears to be "MORE ADVANCED" (higher number). Apple must have added some secret performance boost to those extra space...
iPhone is dumb hardware, upon which is a licenced OS which you are licenced to use (not to OWN the IP to). Without their hardware, you can't use iOS on any other device. Detailing to such ridiculous levels of pedantry is not something which would behoove Apple OR its customers; what would the average man DO with that information anyway? They don't have any need to reveal it, because it starts stupid threads like this one, where every Joe Schmoe and his uncle jump on their uninformed high horse and create problems which aren't even there, and wouldn't even understand the difference in parts if they DID KNOW.
Do you not think Apple may just have engineers who know a littttttttle bit more about it than a bunch of randoms on t'internet? I reckon they do
Wow, I feel a major slapping of palm to face, this is one of those "Dumb as F***" threads.
There's a 70 percent chance you have a Samsung chip, and a 30 percent chance you have a TSMC chip. There's a 100 percent chance you have the fastest phone on the market currently. Hope that helps.
Agreed, I don't care which one I would end up with, but I would really want to know just to have that little piece of knowledge.
I misspoke when I mentioned architecture. I was thinking of instruction sets, etc. Good catch.