...
I have not suggested that the current A-series are a good fit for MacOS devices.
So you want theoretical numbers for a processor that nobody has any physical baseline implementations of? You want a projection on something that Apple is
not doing. Hasn't planned to do. And actually has hinted is not a priority at all.
' ...
And since Apple is doing a fine job with mobile processors, it could conceivably decide to get into conventional chips and bump Intel out of its Mac laptops and desktops. Srouji, of course, won’t go there, though he does allow that his team’s mission is finite. “If we attempt to do everything on the planet,” he says, “I don’t think that would be very smart.” ... '
http://www.bloomberg.com/features/2016-johny-srouji-apple-chief-chipmaker/
If you want numbers pulled out of your butt, then just make them up. Take the benchmark, divide by GHz, and then multiply by whatever GHz you want to make up that is higher. Then take that and multiply by 4. Linear scaling is a theory you can use. It has not too much to do at all with theory of actually modeling computer architecture implementations, but it works as a "pull numbers out of my butt" mechanism.
I read an article one or two years ago, which stated laptops make up about two-thirds of Apple's computer sales. So we are already there.
No we are not. The MacBook is a product. I was not using 'Macbook' as a generalized concept in my quote. It was directed at that specifically named product. It is the lowest "horsepower" Mac in Apple's line up. That could be bumped to ARM. However, it would be shocking if that was even 25% of Apple's laptop sales. It is no where near 2/3.
Maybe if throw in the MacBook Air (which plays the role the older MacBook's used to play as being the "most affordable" laptop in the Mac Line up.) you might climb with in the vicinity 50%. However, ARM does nothing for the Macbook Pro, so there is no way getting anywhere getting close to 2/3.
It remains to be seen how much the iPad Pros are going to eat into that lower edge of the MBA and the very low workload, but highly focused on weight subset of the MacBook market. If anything should be trying to get the MBA/Macbook to move higher out of the way of the iPad Pro if Pros start significant cannibalization.
It seems repetition is the best way to get my point across, so again, I'm not saying the A-series are suitable for Macs. I'm merely speculating about Apple's ability to design processors that are.
Designing phones processors doesn't multiply the team size. They aren't rabbits in a hutch inside of Apple and just multiply if keep them feed with food. They have a team very busy with phones, watches , and now in part, earphones. The number of custom chips inside of the phones is more likely to go up over time. If there is team expansion it is far more likely it will get consumed by that rather than a distraction of Macs.
In the x86 market ( if AMD's Zen pans out as early numbers indicate), they have two competitors vying to supply Mac processors. As long as they are doing a good job.... what is the point?
If Apple has mega resources of engineering talent just lying around doing nothing, where are the new Macs? There is a whole line up of nothing new. Apple does appear to have the ability to throw bodies at empty "green field" lines of development ( 1,000 folks on car , 100 folks on watch bands in the initial ramp up, etc.) but longer mature/maturing product lines Apple seems to pick out a "this is big enough" size and then attempts to deal with whatever that approximately sized group can do.
The phone processor is extremely critical to Apple maintaining their level of penetration in the "expensive phone" market. If Apple lets Qualcomm , Samsung , or some newcomer blown past them they'll be in bad shape on trying to create differentiation. It is highly doubtful Apple will take on Cellular radios so that key component will be shared.
Yes. I'll agree Gruber dropped the ball here (to the point where it looks like he purposely cherry-picked his data to make the A10 look good). But even so, I do think the Geekbench illustrates just how far Apple's processors have come in a relatively short time.
Implementing stuff that was deeply researched and documented 10-15 years ago. Is going to run out of steam at same walls as the other stuff that has already implemented it 5-6 years ago.
The pace at which ARM improves is reminiscent of the leaps x86 made in the late 1990's.
There is little in these A10 numbers but Megahertz. To a large extent Apple is just simply clocking a single core higher than their competitors on single core "drag racing" benchmarks. As soon as get to multicore work the advantage largely disappears. So as long as have a workload that using
one of two cores they run much faster. That is not all that amazing.
It will be interesting to see what happens when folks throw single core workloads at the A10 that last for 10-30 mins. ( not the 1-2 geekbench windows). I highly suspect there isn't going to be some huge gap that A10 has at that point and that the Intel/AMD implementations used in most of the Mac laptop line up leave them far behind.
