Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

pgiguere1

macrumors 68020
Original poster
May 28, 2009
2,172
1,247
Montreal, Canada
Anand Shimpi from Anandtech has been pretty confident since the iPhone 5's announcement that the A6 uses Cortex A15 cores, but it's been unclear why. This guy is known to provide pretty accurate analysis/predictions and doesn't usually just state speculation as facts unless he's sure of his claims, and as you may already know, Anand knows his stuff.

Turns out Anand apparently has some exclusive info on the A6, judging by the latest tweets on his personal Twitter account (@anandshimpi).

anand.png


He has some proofs the A6 uses Cortex A15 cores, but they'll have to wait until later.

The other two possibilities people considered to get such a performance gain with the A6 over the A5 were:

1) The A6 is a 32nm A5 with a much higher clock speed
2) The A6 is a 32nm quad-core A5

The first one is completely improbable given the fact that the iPhone 5 has a better battery life than the 4S even with an only marginally better battery capacity and the use of a larger display and LTE. To have an A5 running at around 1.6GHz would completely destroy battery life, even with a 32nm die shrink. Also, notice how Apple themselves are saying that the A6 is more efficient than the A5:

46626484.png

efficiency.png


An A5 overclocked by 100% is certainly not an efficient solution.

The second possibility was also pretty improbable just from a marketing standpoint. We all know Apple would have made a big deal out of quad-core, just like they made a big deal out of the A5's two cores and the A5X's 4 GPU cores. But if that wasn't enough of a proof, Anand confirmed it wasn't technically possible either following Apple's speed increase benchmarks:

threaded.png


The specific tasks Apple cited to compare the speed increase over the 4S's A5 would have insignificant speed gains due to the inclusion of additional cores at a similar clock speed, according to Anand, since they aren't tasks that take real advantage of multi-threading.

iPhone5-301_575px.jpg


Those are however not 100% solid proofs, and we can't make a final conclusion with the details we currently know, but it seems like Anand would get in trouble if he revealed more of what he knows, whatever it is. This is still pretty much as close of a proof that the A6 is Cortex A15-powered that we'll get until the inevitable teardown following the iPhone 5's release.
 
So is this news GOOD news?

I have no idea what that means.

Apple is going to be the first company to ship a product with a chip in it (A15) which has smoke absolutely every competitor, even the 1.4Ghz chip in the Samsung Galaxy Note.
 
So is this news GOOD news?

I have no idea what that means.

Yes, Cortex A15 is the very best CPU architecture we could hope for, and that would make the iPhone 5 the first device to use it and (probably) the most powerful mobile device on the market for a while.

Considering that iOS is pretty resource-efficient and that iPhones usually have better real-world performances for similar specs compared to competing smartphones, that would make the iPhone 5 a real beast.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.