someone (someone.delete@this.somewhere.com) on September 23, 2013 9:43 am wrote:
> jose (1.delete@this.2.com) on September 23, 2013 4:43 am wrote:
> > A while ago Charlie Demerjian was almost eaten alive when he stated that
> > Apple wanted/was planning to migrate to ARM for their laptop/desktop products.
> > Is it a coincidence that Apple is the first one with an ARM64 chip?
> >
> > Regards,
> >
> > Jose
> >
>
> The latest Macbook air with Haswell's single thread performance and multithread
> throughput measured with Geekbench 3 (sigh) are both more than double that of
> an A7 running 64b and triple A7 running 32b.
>
> Apple has a *long* way to go before it can replace Intel in its Macs without a huge
> drop in performance.

You are missing the big picture.
This claim WAS a crazy claim even with the A6 --- the gap between A6 and say IB was just too large.
With the A7 the gap no longer exists at the µArch level --- we have equal IPC. Now it's purely a question of implementation. Could Apple (if they wanted) fab much the same design at 2 to 3x the frequency?
We know Krait runs at 2.1GHz, so the answer is they probably could, if they wanted, get to 2x the frequency. More than that is tough. When TSMC hit 20nm?

The thing to remember is that the business case for this is HUGE. Yes, Apple sell more iPhones. But the cost of ARMs, if they bought them from Qualcomm or whatever, is nothing. Apple designs their own as much to control the future as to save money.
With Intel, the numbers are very different. A nice Intel CPU costs you what, $350 or so --- if Apple makes their own, even with the other costs that come along, that allows for a substantial drop in Apple's prices for MBAs and mac mini's .

The point is, there is probably enough money involved that you can afford to employ an additional design team to weak the design for performance rather than power, and to create a new set of masks. AND you get all the advantages of controlling your own destiny, making the tradeoffs YOU want.

This is not the same thing as claiming iOS on the desktop. That's crazy --- all Apple needs to do is look at Win8 to see what a bad idea that is. But Apple has been happy to switch CPUs before and, let's face it, OBVIOUSLY they've already built OSX for ARM-64 inside Infinite Loop --- I mean, is there a single person who doesn't expect that? (Have they also built a x64-on-ARM-64 emulator? That's more work, and really shows their intentions. Probably, and so THAT would be the interesting thing to know...)

They'll pay a small price in user base by losing Windows compatibility, but they gain a heck of a lot too.

If the best they can do with Cyclone is to match Intel, it's a finely balanced case and may not be worth it. But Cyclone is only rev two. If they believe there is a whole lot more oomph there, another 50% in IPC available just by ditching the 32-bit bwd compatibility stuff, that changes the calculation a lot.

To match Intel on the desktop they would need to add
- GPU (probably coming)
- HWTM (my guess is this is coming)
- a truly kickass memory controller (They may have this already --- I've seen no discussion of the quality of the A7 memory controller compared to Intel)
- (probably) PCIe support stuff (this they probably don't have, but can live with some 3rd party stuff that runs at 70% of Intel performance)