They also have the option of AMD Zen's that are coming to keep some machines running X86 platform since they tend to stick with AMD GPUs.
I'd think that the threat of Zen be a bargaining chip for Apple to get really good deals. I recall reading something recently that Apple had secured "really good" deals for Skylake chips. I am taking this to mean Skylake where Kaby Lake equivalents haven't been made readily available by now.
To secure USB-C / Thunderbolt 3 compatibility they just need to implement the Alpine Ridge controller chip which just adds a little bit to the final bill of materials on stuff like the iMac. It just takes up a bit more room on the motherboard and might make a bit more heat/be more power hungry than Apple would like.
Mark Gurman's Mac upgrades and refresh article on Bloomberg at the back end of August
specifically mentions a Macbook Air. The bit about the iPad has come true - with no refresh till next year.
Given that earlier this year the 11" MBA was left untouched and the 13" MBA only got a RAM bump to 8Gb base for all models, I had assumed that the MBA would get repurposed as the EOL cheapest Macbook option in advance of being retired while the retina Macbook Pros got updated with Kaby Lake.
It may still do when the refresh comes in a matter of weeks unless Apple decide that the recently departed Ivy Bridge non retina Mac model will be in fact replaced by a poverty spec Broadwell based 13" retina Macbook Pro for the next 4 years instead.
The 2014 Mac Mini model relies on Intel's 28w U series part which isn't being updated to Kaby Lake until 2017. The part should go in the retina Macbook Pro 13". This is a similar story to the retina Macbook Pro 15" models which rely on the 45w HQ series parts which are similarly not being updated until 2017.
I would contend that if Apple are going to go USB-C/Thunderbolt 3 with all their Macs this year they can do it the following way:
Retina Macbook refresh to use Kaby Lake 4.5w CPU (native USB-C/Thunderbolt 3)
Retina Macbook Air to use Kaby Lake 15w CPU (native USB-C/Thunderbolt 3)
Retina Macbook Pro 13" gets dropped to bargain basement model
Retina Macbook Pro 15" gets HQ series Skylake; i5-6350H quad option lowers base model starting price, while upper models get AMD Polaris graphics. We therefore get 3 models of the 15" to choose from - base i5, base i7 and i7 with Polaris.
For information, the iMac 21.5" would use the Skylake-H desktop*, with at least middle model going retina, and using Iris Pro 580 graphics.
And the 27" model still uses Skylake chips but with AMD Polaris options and USB-C/Thunderbolt 3 ports.
* The base 21.5" iMac has been non retina and using Macbook Air CPU, which was the same arrangement as the base model Mac Mini, but this could be in danger if Apple want to go fully retina across the range including the iMacs.
Alright, so what has all this got to do with the Mac Mini?
Well, the 2 things we know about the Mini are it's marketed as Apple's most energy efficient desktop computer, and it's typically been based on the 13" retina Macbook Pro in terms of internals.
So despite the cost effective option of the i5-6350H I think Apple would actually use the 15w Kaby Lake CPUs and continue the line of most power efficient desktop computer but introduce it across the range. (yes, gasps from Mac Mini fans). I can see performance per watt coming into the inevitable press release as the Mini won't get an on stage intro.
This is the year in which a 1Tb Fusion drive could become standard on Mac mini to make it feel like a more responsive computer, and if Apple wanted to be cost effective, they could probably carry on using the same case of the 2014 Mac Mini and just redesign the ports on the back which are in effect on the motherboard. With such massive extra thermal head room a 2016 Mini might be able to stay in turbo boost mode for longer before the fans kick in.
[doublepost=1474144025][/doublepost]
This seems like the right point in the thread to post this excerpt from an article on The Verge
http://www.theverge.com/2016/9/16/12939310/iphone-7-a10-fusion-processor-apple-intel-future
"Intel already let Apple down once with the
original MacBook’s underpowered Core M’s underpowered Core M processor, and the absence of an Intel Skylake upgrade for either the MacBook Pro or Air this year also seems to have been caused by Apple’s dissatisfaction with Intel’s CPUs. There is nothing that Apple would like to do more than rid itself of its reliance on Intel, which would eliminate such unforeseen hiccups in the future. Apple is famous for its
vertical integration, owning and controlling every possible aspect of its supply and production chain, and switching to its own processor chips across all devices would be the next logical improvement in that integration.
If you want to develop the next great processor, you’d better be going mobile first and building from there. Intel’s decades of futile attempts to
shrink desktop chips into mobile devices have shown how not to do things. Now Apple’s persistent and apparently accelerating performance improvement with the A series suggests it might have found the right path through. It’s fitting that this new chip is called A10 Fusion, because the path it’s leading us on will eventually lead to the merging of what we now consider two distinct classes of mobile and desktop PCs".
The A series chips are powerful yes, but I'd have thought Apple would only use a version of iOS to run it with a fully locked down app store. It'd be nice if they actually looked seriously at the likes of QNAP and Synology and reinvented an Apple Server like some of the 4+ bay professional model NAS drives you see on sale.
Kind of like a Time Capsule Pro.
And if it could use iSCSI and plug into Macs like professional external storage that would be even better. But for a general purpose computer with A series CPU - they call that the iPad Pro.
