Apple A8 and A8X

[AlNom]

Couldn't they have at least added the anti-reflective coating? Not sure how much that really adds to the cost, but it'd be a bit more attractive for marketing to the clueless (IMHO).

 

[silent_guy]

It's interesting that the iPad Mini 3 uses the A7 and supports in app Apple Pay while the iPhone 5s doesn't.

The 5s won't have Apple Pay for websites and apps?

 

[Lazy8s]

I hadn't stopped to think about it before my last reply, but it's probably just a case of them not highlighting it at the September event like they did for the new iPads at yesterday's Octobet event.

Would they really go out of their way to stop any Touch ID device from using online Apple Pay? I imagine the feature is there and working, just not mentioned.

 

[ltcommander.data]

The 5s won't have Apple Pay for websites and apps?

https://www.apple.com/apple-pay/

Apple has a compatibility chart at the bottom of their Apple Pay page. The new iPads can do in app only, the Watch can do in store only, and the iPhone 6/6 Plus can do both. The iPhone 5s is not mentioned so I assume it's not supported.

I hadn't stopped to think about it before my last reply, but it's probably just a case of them not highlighting it at the September event like they did for the new iPads at yesterday's Octobet event.

Would they really go out of their way to stop any Touch ID device from using online Apple Pay? I imagine the feature is there and working, just not mentioned.

I agree that at this early stage Apple should be trying to maximize the Apple Pay installed base to make an attractive target for developers/retailers. So if the iPhone 5s doesn't support even in app Apple Pay I'm assuming there is some technical limitation.

 

[patsu]

Well presumably, after they get the Apple Pay thing going, they could use the NFC for other things.

They could open it up to third-parties, the way they opened up TouchID after one year.

So it could have other uses.

May be they want people to buy AppleWatch ?

Personally, I don't think Apple Pay, or rather Passbook, is limited to NFC.
IMHO, there are plenty of use cases for pay-anywhere.

We probably are just looking at the first implementation of Apple Pay.

 

[wco81]

My understanding is that the SE is a part of the SIM card and thus, on Android devices, carriers control access to it or simply disable it, which hobbles mobile payment efforts.

So they use HCE which stores virtual cards in the cloud?

http://blog.simplytapp.com/2014/09/apple-pay-and-android-payment-eco.html

Apple put in dedicated HCE in the iPhone 6 and 6+ and they alone control it for now.

Plus the Apple SIM virtual SIM tech may also play into the role of the SE on the SIM.

 

[patsu]

That's just the offline NFC payment flow. We know Apple Pay also supports online payment (by the apps).

If they go to China or other countries, they should be able to extend the payment family to include the respective national payment infrastructures (whatever they are).

 

[pcchen]

Apple Pay's online payment system is likely to be very similar to its offline payment process. Basically, the app request the system for the amount of payment, the merchant ID, etc. and the system generates an one-time use token from the card data in the secure enclave. Then the data is sent to the merchant's server, which then sends it to Apple's server for validation. Once it's validated, the merchant can safely deliver the merchandise.

So for on-line payment there are two essential elements: the secure enclave and Touch ID. Since iPhone 5S have Touch ID, I think maybe its secure enclave is not prepared for supporting Apple Pay. Of course, it could also be just an artificial limitation.

 

[wco81]

I recall reading somewhere that online shopping occurs more in devices with bigger screens.

So more on iPad and tablets than smart phones and presumably more on desktops than mobile devices.

So online shopping from mobiles is a fraction of that from desktops. That may change in 5-10 years and we see that online shopping is growing relative to retail shopping.

Mobile payment systems may not play as big a role in online shopping then.

 

[Grall]

You people sure ipay don't work on iP5S? Because, well, it does in the app store. So if it's good enough there, why not elsewhere as well...?

 

[wco81]

I think they want to nudge people to upgrade.

They're running some relatively generous trade-ins on the 5S.

If I can find an unlocked Plus, I might do the trade in.

 

[Grall]

Yeah, you're probably right. In any case, it doesn't really matter, as I'm going to get myself an iPad this year, so it'll be the Air2, and next year iWatch comes out, so that should cover me anyway I think...

The 5S is such a great phone as-is, I really don't see the pressing urge to upgrade. It'd just be money wasted really, as I'd have to end my contract early, and I can't really afford that. Better to let it run its full two years and then upgrade; the upgrade will be more substantial then as well, with Apple hopefully using TSMC 16nm FF+ process by then for the SoC.

 

[iMacmatician]

An iPad5,4 (presumably the iPad Air 2?) has appeared in PassMark's charts. According to the CPU Mark chart it has 3 CPU cores and the memory chart seems to show 2 GB RAM (in both cases, mouse over the bar corresponding to the iPad5,4).

Is this information reliable?

 

[ams]

An iPad5,4 (presumably the iPad Air 2?) has appeared in PassMark's charts. According to the CPU Mark chart it has 3 CPU cores and the memory chart seems to show 2 GB RAM (in both cases, mouse over the bar corresponding to the iPad5,4).

Is this information reliable?

Well that is certainly interesting and could help explain in part why A8X has 1 billion more transistors than A8!

 

[Ailuros]

It needs a lot more than one additonal CPU core to explain that almost +1b transistors. I hope that the +40% CPU performance Apple claimed between iPadAir and iPadAir2 aren't at 1.4GHz for both

 

[ams]

Well A8X appears to have 50% more CPU cores AND 50% more GPU cores than A8, so that would explain the 50% extra transistors.

It also appears that the CPU cores will have a slightly higher clock operating frequency than A8 (say, ~ 1.6GHz?).

 

[Ailuros]

Well A8X appears to have 50% more CPU cores AND 50% more GPU cores than A8, so that would explain the 50% extra transistors.

While scaling ONLY clusters on a GPU, transistor count/die area will not scale in a linear fashion.

It also appears that the CPU cores will have a slightly higher clock operating frequency than A8 (say, ~ 1.6GHz?).

It would have to be higher to make the +40% claim more realistic.

 

[ams]

So the multi-core CPU performance of A8X should be close to 2x better than A7 (while the single core CPU performance should be 1.4x better).

That means a Geekbench 3 AArch64 score of ~ 2000 for single core and ~ 5400 for multi-core!

 

[Ailuros]

I didn't mean synthetic benchmarks. It's good enough for me if the iPad Air2 can show in several spots in real time 40% higher CPU performance compared to iPad Air.

 

[ams]

I agree that the single-threaded CPU performance is far more important (and it is unclear how many mobile apps are actually able to take advantage of more than two CPU cores), but this is a great answer to the quad-core and octa-core Cortex-powered SoC's that promote multi-core benchmark scores.