When Apple announced the iPhone 5, Phil Schiller officially announced what had leaked several days earlier: the phone is powered by Apple's new A6 SoC.

As always, Apple didn't announce clock speeds, CPU microarchitecture, memory bandwidth or GPU details. It did however give us an indication of expected CPU performance:
 
 
Prior to the announcement we speculated the iPhone 5's SoC would simply be a higher clocked version of the 32nm A5r2 used in the iPad 2,4. After all, Apple seems to like saving major architecture shifts for the iPad. 
 
However, just prior to the announcement I received some information pointing to a move away from the ARM Cortex A9 used in the A5. Given Apple's reliance on fully licensed ARM cores in the past, the expected performance gains and unpublishable information that started all of this I concluded Apple's A6 SoC likely featured two ARM Cortex A15 cores. 
 
It turns out I was wrong. But pleasantly surprised.
 
The A6 is the first Apple SoC to use its own ARMv7 based processor design. The CPU core(s) aren't based on a vanilla A9 or A15 design from ARM IP, but instead are something of Apple's own creation.
 

Hints in Xcode 4.5

 
The iPhone 5 will ship with and only run iOS 6.0. To coincide with the launch of iOS 6.0, Apple has seeded developers with a newer version of its development tools. Xcode 4.5 makes two major changes: it drops support for the ARMv6 ISA (used by the ARM11 core in the iPhone 2G and iPhone 3G), keeps support for ARMv7 (used by modern ARM cores) and it adds support for a new architecture target designed to support the new A6 SoC: armv7s.
 

 
What's the main difference between the armv7 and armv7s architecture targets for the LLVM C compiler? The presence of VFPv4 support. The armv7s target supports it, the v7 target doesn't. Why does this matter?
 
Only the Cortex A5, A7 and A15 support the VFPv4 extensions to the ARMv7-A ISA. The Cortex A8 and A9 top out at VFPv3. If you want to get really specific, the Cortex A5 and A7 implement a 16 register VFPv4 FPU, while the A15 features a 32 register implementation. The point is, if your architecture supports VFPv4 then it isn't a Cortex A8 or A9.
 
It's pretty easy to dismiss the A5 and A7 as neither of those architectures is significantly faster than the Cortex A9 used in Apple's A5. The obvious conclusion then is Apple implemented a pair of A15s in its A6 SoC.
 
For unpublishable reasons, I knew the A6 SoC wasn't based on ARM's Cortex A9, but I immediately assumed that the only other option was the Cortex A15. I foolishly cast aside the other major possibility: an Apple developed ARMv7 processor core.
 

Balancing Battery Life and Performance

 
There are two types of ARM licensees: those who license a specific processor core (e.g. Cortex A8, A9, A15), and those who license an ARM instruction set architecture for custom implementation (e.g. ARMv7 ISA). For a long time it's been known that Apple has both types of licenses. Qualcomm is in a similar situation; it licenses individual ARM cores for use in some SoCs (e.g. the MSM8x25/Snapdragon S4 Play uses ARM Cortex A5s) as well as licenses the ARM instruction set for use by its own processors (e.g. Scorpion/Krait implement in the ARMv7 ISA).
 
For a while now I'd heard that Apple was working on its own ARM based CPU core, but last I heard Apple was having issues making it work. I assumed that it was too early for Apple's own design to be ready. It turns out that it's not. Based on a lot of digging over the past couple of days, and conversations with the right people, I've confirmed that Apple's A6 SoC is based on Apple's own ARM based CPU core and not the Cortex A15.
 
Implementing VFPv4 tells us that this isn't simply another Cortex A9 design targeted at higher clocks. If I had to guess, I would assume Apple did something similar to Qualcomm this generation: go wider without going substantially deeper. Remember Qualcomm moved from a dual-issue mostly in-order architecture to a three-wide out-of-order machine with Krait. ARM went from two-wide OoO to three-wide OoO but in the process also heavily pursued clock speed by dramatically increasing the depth of the machine.
 
The deeper machine plus much wider front end and execution engines drives both power and performance up. Rumor has it that the original design goal for ARM's Cortex A15 was servers, and it's only through big.LITTLE (or other clever techniques) that the A15 would be suitable for smartphones. Given Apple's intense focus on power consumption, skipping the A15 would make sense but performance still had to improve.

Why not just run the Cortex A9 cores from Apple's A5 at higher frequencies? It's tempting, after all that's what many others have done in the space, but sub-optimal from a design perspective. As we learned during the Pentium 4 days, simply relying on frequency scaling to deliver generational performance improvements results in reduced power efficiency over the long run. 
 
To push frequency you have to push voltage, which has an exponential impact on power consumption. Running your cores as close as possible to their minimum voltage is ideal for battery life. The right approach to scaling CPU performance is a combination of increasing architectural efficiency (instructions executed per clock goes up), multithreading and conservative frequency scaling. Remember that in 2005 Intel hit 3.73GHz with the Pentium Extreme Edition. Seven years later Intel's fastest client CPU only runs at 3.5GHz (3.9GHz with turbo) but has four times the cores and up to 3x the single threaded performance. Architecture, not just frequency, must improve over time.
 
At its keynote, Apple promised longer battery life and 2x better CPU performance. It's clear that the A6 moved to 32nm but it's impossible to extract 2x better performance from the same CPU architecture while improving battery life over only a single process node shrink.
 
Despite all of this, had it not been for some external confirmation, I would've probably settled on a pair of higher clocked A9s as the likely option for the A6. In fact, higher clocked A9s was what we originally claimed would be in the iPhone 5 in our NFC post.
 
I should probably give Apple's CPU team more credit in the future.
 
The bad news is I have no details on the design of Apple's custom core. Despite Apple's willingness to spend on die area, I believe an A15/Krait class CPU core is a likely target. Slightly wider front end, more execution resources, more flexible OoO execution engine, deeper buffers, bigger windows, etc... Support for VFPv4 guarantees a bigger core size than the Cortex A9, it only makes sense that Apple would push the envelope everywhere else as well. I'm particularly interested in frequency targets and whether there's any clever dynamic clock work happening. Someone needs to run Geekbench on an iPhone 5 pronto.
 
I also have no indication how many cores there are. I am assuming two but Apple was careful not to report core count (as it has in the past). We'll get more details as we get our hands on devices in a week. I'm really interested to see what happens once Chipworks and UBM go to town on the A6.
The A6 GPU: PowerVR SGX 543MP3?
POST A COMMENT

162 Comments

View All Comments

  • vision33r - Saturday, September 15, 2012 - link

    At the end of the day majority of the iPhone users buy it for it's simplicity and usefulness without having to spend a ton of swipes to make things work.

    That's the beauty of simplicity that you get things done without having to use a lot of resource.

    Android is a very stubborn OS because it has to cater to those who use it for more than just a simple smartphone. That's why a bigger screen matters to Android users.

    If I were to use an iPhone I would prefer one as small as my watch. I really wish SIRI can do everything for me without me ever lifting a finger.
    Reply
  • doobydoo - Sunday, September 16, 2012 - link

    There's no useful end use that you can't achieve on iOS that you can on Android.

    There are just guys who think they are tech geeks who haven't worked out how to do something on an iPhone.
    Reply
  • solomonshv - Sunday, September 16, 2012 - link

    "Imagine if every time you bought a new car the controls were in a different place. Sometimes the brake is on the left, sometimes it's on the right; sometimes it's on the steering wheel, sometimes on the armrest."

    stupid example, should have farted out a better one.

    first, you went a little too far with that. that's like saying you bought a phone where the microphone is above the screen and the speaker is below that.

    second, it seems to me like you have only owned 1 car in your entire life. i'm on my third SUV from the same japanese maker. things like cruise control, wind shield wiper controls, audio controls always seem to move around.

    hell, even the controls for the transfer case changed from a lever to a knob and moved about 8 inches further away from the gear stick. .
    Reply
  • solomonshv - Sunday, September 16, 2012 - link

    -"Do you need live widgets on there?"
    Yes.

    -"Well I'd argue most people don't."
    Who died and made you god? i'm sure they will decide for themselves when they see widgets in action. You should see the reactions on iPhone/BB user's faces when they play with my GS3 and see the gmail and calendar widgets, and the disappointment that follows when i tell them that only android does that.

    -"Apple iOS is already well established."
    You mean well advertised. The Apple logo is flashed in the viewer eyes in movies, CW soap operas, rap videos, bill boards, etc. They work really hard to put that Apple in the center of the screen. no other company advertises anything so aggressively. when was the last time you saw the Toyota symbol flashed in your face every 5 minutes in a TV show? Characters on prime time shows like Vampire diaries and gossip girls always walk around with the Apple products and the Apple logo appears on the screen so much that it may as well be the protagonist.

    go to youtube, enter "First Look: iPhone 5 jimmy kimmel" in the search box, hit enter and open the first video. (don't click the ads above the actual video).

    -"Do you yell at Windows 7 and wish it was more evolved?"
    No. Windows 7 has all the functionality anyone could possibly desire. thanks to the registry editor, there isn't a single function or menu that can't be modified, removed or added. and the instructions for any modification are readily available to anyone who knows how to use google.

    -"So again, it would have to be a whole new redesign, and again don't think that's warranted whatsoever."
    Of course not. Steve took a zillion surveys that demonstrated that his method was perfect.

    -"It's all about the apps that will round out the OS"
    how about removable storage? and the ability to add songs to your music library on the go without being changed to iTunes and the horrible quality AAC? Maybe a notification LED in the front (i know you can use iPhone's flash as an indicator light, but it is way too bright and will burn the entire battery in a few hours).
    Reply
  • doobydoo - Sunday, September 16, 2012 - link

    Widgets may impress as an immediate gimmick, but in day to day use it's an irrelevance, which simply drains battery life.

    iOS isn't advertised at all, really. Samsung spends far more on marketing than Apple does and when Apple DOES advertise it's for product features - not iOS itself. It's unarguable that iOS is better established than Android - it's plain obvious given Apples consistency and consistent support of legacy devices.

    Jimmy Kimmel youtube video is funny - but proves nothing. Take any modern smartphone and you'll be able to capture similar responses. They no doubt edited out the people who said 'it looks the same'. It's a measure of American stupidity - not stupidity specific to Apple.

    Rants on Windows 7 don't belong on a smartphone discussion.

    As for removable storage - that's hardware, not software - so it's nothing to do with iOS. You can buy an adapter for a couple of dollars to allow you to read SD cards on your iPhone - most don't, because internal storage plus cloud = plenty. SD card storage is slow to read.

    You don't need iTunes, full stop, to do anything - especially not to manage music - you just seem to not have a clue about that.

    Seems like you don't really have a clue what iOS is like these days - you're living in the past while the rest of the world has moved on.
    Reply
  • EnzoFX - Sunday, September 16, 2012 - link

    First you misunderstood completely. Second your argument is contradictory and not to mention rather ignorant, shows the blind bias really. Lastly all it goes to prove is that you are not the 99% that the phone is catered to. Failing to realize this is the most ignorant thing you can do when trying to make any comparison argument here.

    You couldn't even acknowledge that iOS is well established lol. Let alone realize its the better app scene.
    Reply
  • Scannall - Sunday, September 16, 2012 - link

    How exactly is iOS a piece of crap? It's a lot more efficient and faster than Android. Runs apps better and faster. Is a lot better at multi-core support. Less likely to have battery sucking apps in the background. Has a much better SDK. A better eco-system. Runs smoother and always has.

    Is it because it looks the same? Windows has had pretty much the same look since Windows 95. Just a slow evolution. Yet, with Windows 8 coming, and a drastic change you have people screaming from the rooftops over it.
    Reply
  • Penti - Saturday, September 15, 2012 - link

    It doesn't really mean it's not A9 does it? It might only mean that the Old Intrinsity team and Samsung in Texas designed a A9-based chip from synthesizable RTL, changing out the FPU. It does however mean that they don't use any hard-macros or such from Samsung though, but they are still probably involved. Interesting to see either way especially if it's their own custom ARM-compatible processor but I doubt there is any performance enhancements in the pipeline except for the added VFPv4.

    Remember even the A5X had roughly double the GPU power of the A5 or previous smartphone / iPhone SoC's/processors. It's not twice as fast gpu as A5X here.

    If it is their own architecture actually then it's the old PA-Semi team that has designed it probably, not anyone in Texas, but if so why was it so adamant to buy the Intrinsity team that designed the A4 (and Hummingbird) with Samsung in Texas?

    32nm mean at least that it is not Krait. But do we really know who's involved here? 28nm and another gpu vendor would mean it is a rebadged Qualcomm chip, but shouldn't the drivers leave any marks if so?

    You at least forgetting there are three ways to implement ARM, one to design your own ARMv7/8 compatible architecture, two to implement and adapt a synthesizable Cortex-core from ARM which means you can adapt it for certain processes, power enhancements, fabrication and so on, or three to license a hard-macro from ARM/the fab and combine that with other logic IP.

    As VFPv3 is optional (or at least configurable Tegra 2 with VFPv3-D16 is an example) on the Cortex A9 it's not totally alien to fathom that they also could replace it if they work with ARM here or work with say Samsung's fab which has a full license for ARM cores. Any way they seem to intentionally be very quiet about the subject.
    Reply
  • ltcommanderdata - Saturday, September 15, 2012 - link

    "If it is their own architecture actually then it's the old PA-Semi team that has designed it probably, not anyone in Texas, but if so why was it so adamant to buy the Intrinsity team that designed the A4 (and Hummingbird) with Samsung in Texas?"

    PA-Semi's contribution would likely be on the architecture while Intrinsity's specialty is in laying out the transistors to maximize performance and lower power consumption.
    Reply
  • Freakie - Sunday, September 16, 2012 - link

    I agree that there is probably more here than even Anand is acknowledging could be (also Anand seems to be trying to throw as much good light as he can at Apple for this). Apple already has used ARM ISA designs in the past, and even went further and tweaked the Samsung's to make it just a bit more like how they want (or rather, PA Semi did that I believe... highly doubt there are any engineers actually at Apple that can do that stuff when they already own companies that do it). I doubt that this situation is any different than that previous situation, where they have just taken an already customized package and customized it some more. Of course, I think we wont know that for-sure until we get die-shots of the thing, like how we were able to tell that their previous one was almost exactly a Samsung processor. Reply

Log in

Don't have an account? Sign up now