Apple’s powerful A6 a unique CPU design that’s never been seen before

Microprocessor Report’s Linley Gwennap today pens a rather fascinating article about the ‘A6’ chip in Apple’s iPhone 5,” Tiernan Ray reports for Barron’s. “Gwennap writes that Apple has developed its own central processing unit, or CPU ‘core,’ based on the chip design technology of ARM Holdings, but completely different from the chips based on ARM intellectual property developed by Qualcomm, Nvidia, and Texas Instruments.”

Ray reports, “Writes Gwennap, ‘At the iPhone 5 announcement, Apple did not disclose what CPU it used in its new A6 application processor. Recent information, however, indicates that it designed its own CPU rather than licensing a Cortex-A9 or next-generation Cortex-A15 from ARM. Rumors about Apple working on a CPU design have swirled for the past several years, but the company had never publicly confirmed them. Initial analysis of shipping iPhone 5 units, however, confirms that the A6 uses a unique ARM-compatible CPU design that has never been seen before.'”

Apple A6 processor
Apple A6 processor

Read more in the full article here.

[Thanks to MacDailyNews Reader “Fred Mertz” for the heads up.]

Related articles:
Analysis of Apple A6 core reveals exquisite, optimized custom layout done by hand – September 26, 2012
Apple A6 die reveals 3-core GPU – September 21, 2012
iPhone 5′s A6 SoC SunSpider performance fastest ever recorded on a smartphone – September 19, 2012
Apple’s custom A6 processor the result of years of effort, including a $500 million chip development program – September 18, 2012
A6 is Apple’s first with custom-designed CPU cores; iPhone 5 memory size and speed revealed – September 16, 2012

22 Comments

  1. I think we’ll be seeing Apple using increasing numbers of unique, in-house designed parts in future devices.

    It’s pretty obvious that the courts can’t stop rivals from quickly copying Apple’s designs, but from now on, the copies will have to be made with off the shelf parts rather than custom parts optimised for the job.

    1. Unfortunately, for the poor fandroids, Sansung and all android companies have managed to brain wash their fan’s brains (if they have one) so no matter how good apple does, there will be tons of poor bastards brainless androids.. and some microsoft fans, vary few but still some.

    2. Asymco .com talked about the potential downside to this scenario, at least in the future.

      As long as Apple’s iPhone is always *subjectively* better – meaning the last iPhone wasn’t “good enough” and the newest release is deemed “better – then an integrated, Apple-esque approach – integration of parts will be the winner. Clearly the iPhones so far have all been subjectivly better than the previous one.

      The downside is that when a product becomes good enough – AKA PCs and whatnot – any vendor can make a “good enough” product – and the expense of making an integrated product is a liability. Apple can see this happening when the older, 4S or 4 starts selling in higher numbers – the 3Gs sold ~3% when it was the free on contract phone. when Apple sees their older models start selling in volume, thnen we’ve reached “good enough”.

      At that time, we might see Apple abandon it’s own silicon, and go back to using off -the-shelf CPUs. Apple kept a build of Mac OS X for intel CPUs for a long time, for when the PPC ran out of steam.

      I wonder what Apple is doing for the day the iPhone becomes “good enough”.

      1. I believe 99% if what you say is right, but I think that, if we’re seeing a repeat pattern of the PC industry, it’s happening much faster. PC’s were around for decades before you could get one free with a 2-year AOL subscription. With phones, that was already happening when the iPhone came-out.

        The Android makers have already arrived at the same place PC box assemblers — as MDN calls them — are, that whole race to the bottom thing (free, BOGO, etc). My point is, most of these handset makers won’t have the luxury of the super-profits that the PC guys got; with nothing in the bank, they will drop-out fast. “Good enough” will only sustain one (Samsung?) or a few.

      2. One difference is Apple’s enormous scale. In a few years they may be selling many hundreds of millions of CPUs per year. At that scale, and given the benefits a custom chip cold bring it may not be that expensive. Plus there may be some strategic things we don’t know about. Could some hidden circuits be used in securing the device from malware? Maybe there is a split architecture with some memory addressing walled off from Apps? This is potentially useful if the iPhone and iPad will be used as a digital wallet.

        1. Exactly the scale of Apple chip production already is far greater than macs ever were and equally order chips still potentially have a life on continuing older or new lower spec’d devices. meanwhile there are various competor chips within the ARM umbrella so no one company like Intel has dominance and massive scale beyond Apple. No real comparison with the PC world certainly as yet or for some time to come I suspect.

      3. Yes, race to the bottom and commoditization is already happening, much faster than in the PC industry… this is happening in the *phone* industry.

        Difference is, Apple isn’t making phones. It’s making mobile PCs with a phone app.

        When people realize this, Apple doesn’t have a problem selling its products against phones. As good as Android may be, it is largely java and uses java-esque applet type apps, and it just isn’t up to the robustness and capability of iOS which is basically OS X without some of the unnecessary components. Throw in the facts that the OS on your Android never gets updated, and the phone itself is of low quality construction, and the difference is readily apparent — the difference in price is justified by these alone.

        Things like CoreAnimation are just now coming into their own on iOS as the hardware is getting more capable — that very hardware that people are concerned that Apple is over-serving when everything else off the shelf is “good enough”. Trouble is, all Apple’s competitors can do to improve UI smoothness and what-not is to throw more CPU cycles at it. Apple has different approach, and it will pay off.

        Apple is not delivering a phone that over-serves. No, Apple is delivering a miniaturized Mac and non-smart phone consumers and people in the developing world will be well served by them because the iPhone may be their only computer, and with it they can run their whole business, do their marketing, network, create and edit videos, do their finances, plug into all sorts of online services that have great apps with unique user experience…. etc.

      4. I think you misread asymco.

        Make each new generation good enough better to create interest. Keep some of your innovations for the next generation. Improving too much too fast cost more than it returns.

        I doubt there will come a day when the iPhone becomes “good enough” to enough people.

  2. Apple’s strategy is freaking awesome here. First, they can even further differentiate their products (iOS to start, and who’s to say about the future).

    Second, they can make these designs work with the OS, pushing vertical integration to new heights.

    Third, they are still fully compatible with off-the-shelf ARM CPUs, so they are really only risking the money it takes to design these things. If ever they are falling behind in terms of performance they can swap back and not be any further behind the competition. It’s all upside.

  3. Agreed. If Tim’s awesome and perfect for the job it’s for this class of reasons. If Apple are developing proprietary tech in each area and fully vertically integrating, little slips along the way in maps perfection or media spin are negligible.

  4. This guy says Apple is now in an arms race with Qualcomm, etc., a race that is difficult to win.

    Uh, no it’s not difficult to win if you have no intention of sharing your designs with other vendors or selling your chips to other companies.

    Apple is designing specific chips to meet its specific needs for processing power and energy savings. No other smartphone/tablet manufacturer can do that, and this gives Apple a huge advantage.

      1. Not so fast. If the chips start to fail it can send Apple to the bottom of the food chain real quick. Samsung is also developing its own chips as we speak notes they build the A6. Keep on eye on them if you get my drift.

  5. I like the idea of Apple designing its own processors but I wonder if Apple is wasting its time by building its own processors. Android smartphones are using off-the-shelf Tegra 3s and they’re just punching up the clock speed constantly to run faster. The latest RAZR smartphone is running an Atom processor at 2GHz. Consumers probably don’t even know the difference or care about the internals. They just want snappy performance and long battery life.

    You just know the news media is going to put a bad spin on Apple’s proprietary hardware since it makes iOS hardware even more closed which the news media hates. I think it’s best that Apple designs all of its own stuff, so competitors can’t easily duplicate what Apple does. Hopefully the A6 will give Apple a huge cost advantage when producing 50 million units or so. Apple has to think very large scale if its going to continue to expand. Only time will tell.

    1. The media and all the mindless droids can up their clocknspeed to any speed they want, in fact since they are all so infatuated with who’s dick is bigger and so focused on speed, let’s hope they ramp up the speed so high that they run right off the cliff before double checking for smarts.

    2. Don’t worry, they are thinking very large scale.

      Not 50 million, more like 200 million. They could do 50 million A6 devices per quarter between the iPhone 5 and the iPad. The upcoming iPad Mini is purported to use the A6, while the new iPod Touch continues to carry the A5 to differentiate it.

      According to a response on one of the quarterly calls, Apple intends to roughly sell as many new iPhones as all previous iPhones combined. They already sell some 200+ million iOS devices per year. If the iPhone 4S and lower and iPods continue to sell with A5 SoC’s then that raises the margins on those even further.

      Plus, Apple sells one phone for at least a year (the 4 was out about 18 months before the 4S came along). Competitors are always chopping and changing and putting out new flavors every other month in order to get sales.

      So, don’t worry about Apple’s roadmap, they can continue to refine their own chip sets and custom hardware and still make a healthy profit. Apple is through waiting on the chip builder to come up with the necessary technology to meet Apple’s requirements. But Apple can still put the A6 in the iPhone 5S if they had to, perhaps with refinements. Apple has its own in-house chip design team (from PA Semi) and Apple has worked as ARM team member for more than 20 years. They know what they are doing. And they spend less on R&D than anyone else because they are so focused.

    3. Yes but it’s a single core that benchmarks two thirds the performance of apples a6. The 2ghz figure is more of a marketing tool that isn’t providing better performance than current duel and quad core cpu’s in the top phones. That said the atom is a good entry into the mobile market for Intel.

Add Your Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.