Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple's Next-Generation A8 Chip Said to Top 2 GHz, Remain Dual-Core
- Thread starter MacRumors
- Start date
- Sort by reaction score
Performance isn't going to double endlessly because there aren't that many "free" upgrades left anymore. You only move from single to dual core once. You only go from 32-bit to 64-bit once. You only go from off-the-shelf to custom core once. You only go from in-order to out-of-order architecture once.
The thing that raw performance numbers never reveal is power consumption. It's certainly possible to get an A8 or A9 chip to "desktop class" performance (whatever that means, it's a constantly moving target), but the power budget of a phone is going to slap you in the face from trying to run the system at that level for very long.
It's significantly harder to make a legacy app compilable for 64-bit than was the armv6 -> armv7 switch, as long as you didn't mean for example OpenGL ES 1.1 vs. 2. (All armv6-based devices supported 1.1 only, while all armv7-based ones supported 2.0. This, however, has nothing to do with the pure armv6 vs. armv7 difference.)
Actually, I'd say 99.9999% of armv6-compatible legacy apps were armv7-compatible, without any need to change anything in the source (except for, maybe, the non-obligatory deprecation-related changes.) With the 32 -> 64-bit switch, you need to change almost all old for example int declarations, shifts etc. A LOT more change...
We're in the final chapter of the A5 years... iPhone 4S, iPad 2&3, iPad mini.... Apple has been selling A5 devices for going on 5 years when it closes which is "forever" in mobile device and even desktop terms. When they jump to A7/A8 we'll see a massive increase in what devices can do.
Are you defining desktop as an all-in-one or tower PC? Or would you say that laptops are included as well?
See my post above for evidence the A8 could catch current Mac laptops THIS YEAR.
I get that doubling won't happen forever. But Apple has been on dual core since the A4 and only last year went to 64-bit....what accounts for the doubling the other two years?
I'm simply stating that given past evidence its plenty likely we could see another doubling of raw benchmark scores making the A8 comparable to current Intel i5 chips in the MBAs. Doubling the last 4 years hasn't affected battery life noticeably to this point. Moving to a smaller die process coupled with the larger casing and battery means more power efficiency.
This is Apple we're talking about - I think people underestimate all they do in both the hardware and software to optimize performance and efficiency.
I obviously don't know any of this for sure - its speculation like any of the other posts here. But I think I make a pretty good case.
Geekbench is a narrow benchmark to test primary just CPU performance. Systems don't solely consist of CPU performance. The MBA over the last several years are substantially different.
At the low end of the laptop/desktop spectrum there is a bit of an overlap of targeted workload as that of tablets. For the average web browsing, stream video from web , email, chat , tweak photos/videos worlkloads that isn't much CPU differentiation can make. Hence the CPU focus
What is highly dubious is whether can run the whole Mac line up on these completely plateaued workloads. These workloads are just as likely, if not more, likely to be consumed by the iOS devices (or far lower cost Windows devices) than there is any great need to target them with OS X.
Only in geekbench single thread spec porn world. Multiple thread (and concurrent app utilization ) there will remain a substantial difference. As TSX enabled software rolls out it should get a bit better over time.
The A7 is already the most advanced ARM processor in the world. Even the A57 doesn't match it, and nobody is even using it yet.
I'm not sure what the A8 will bring, but if it's a significant upgrade than Apple will be even further ahead of Qualcomm, Samsung and even ARM.
Qualcomm and Samsung need to stop playing the higher clock/moar cores games and start improving their architecture like Apple did with the A7 (and to a lesser extent the A6).
I'm not sure what the A8 will bring, but if it's a significant upgrade than Apple will be even further ahead of Qualcomm, Samsung and even ARM.
Qualcomm and Samsung need to stop playing the higher clock/moar cores games and start improving their architecture like Apple did with the A7 (and to a lesser extent the A6).
I highly doubt you read what I wrote as quad cores in the non myopic sense are already out in the Ax series. For general iOS core OS code the GP cores don't have a high need for more than 2 since can't particular run more than one app at the same time. There is gnerally one human users and one app running.... where is the general high need for 4 cores going to come from?
I agree its not an all encompassing benchmark - but we're comparing two completely different architectures and two different types of software (iOS vs OS X have some very different priorities).
I never claimed that the entire Mac line would run on ARM chips in the next year. Simply that an A8, if we see the same performance increase we've seen over the last 4 years, will be plenty powerful enough to run the base model MBA and potentially Mac Mini. The goal would be maximum battery life for the MBA - while I don't know enough about Intel chip architecture to know what the benefits of an A8 would be in the Mac Mini (a computer that is always plugged in). Perhaps the ability for an almost pocketable desktop?
If this chip will give a 50% boost in power it will go to this range of performance:
http://browser.primatelabs.com/geek...:"Intel Core i5-4250U" frequency:1300 bits:32
But increasing frequency and making the chip smaller with even more transistors will give it up to twice the performance of A7.
Whats the point of using Intel 4.5W low power Broadwell fanless Chips in 12 inch iPad if this one will be more powerful in smaller power envelope?
Of course, IF this is what will happen when Apple will release iPhone and iPads with this chip...
http://browser.primatelabs.com/geek...:"Intel Core i5-4250U" frequency:1300 bits:32
But increasing frequency and making the chip smaller with even more transistors will give it up to twice the performance of A7.
Whats the point of using Intel 4.5W low power Broadwell fanless Chips in 12 inch iPad if this one will be more powerful in smaller power envelope?
Of course, IF this is what will happen when Apple will release iPhone and iPads with this chip...
Thinking of the mobile site I'm developing right now and trying to decide on the best method of optimization is a kicker, because sometimes SVG is better and other time PNG is better, some times you can program something directly with html5 so both should be avoided. Also, I've noticed, you can really optimize SVG codes to reduce the space that they use. When you start integrating media then it starts to really make a difference.
Because it won't be more powerful? Broadwell is an expected 30% increase from the Haswell. That means it should still be more powerful than the A8.
But what happens when we switch from silicon based chips to other types of wafers such as vanadium dioxide.
Then the doubling might start again. I think tech goes through a maturing phase until the next major step change is introduced.I'm not doing either one. I am getting at Apple's hyperbole about "desktop performance" is really not about "desktop" as a general usage term.
Both of those all-in-one and towers are in the desktop class of devices.
In outlining or being representative of desktops? No. As being in the class of "fast enough for most folks" then yes. Same issue of what general usage of laptops lead to decline of desktops as a percentage of what folks buy is same factor as what is now leading to folks passing up both laptops and desktops for tablets. The workloads plateau so CPU speed is relatively a non high value issue.
Laptops are not desktops. So matching bottom end performance laptops really is zero indication are moving into the desktop performance range at all.
The faster the CPU the more versions of iOS it should support. More ram would be nice for backgrounded apps. Especially Safari tabs.
It's not just software. It's the fact that it's a hand held device. You aren't going to do raw photo and video editing very much on an iPhone or iPad so the software isn't taking advantage of it because there is no reason to. Since 95% (I made that number up) of users are just using their phones for texting/ email/ Facebook/ light gaming/ web browsing I just don't see a reason for most developers to ever really unlock the full potential of these new mobile processors. Some surely will but not anything that will be used by the masses.
It is said to be 800 MHz at base and 2.0 GHz in Turbo Mode. MBA's have higher frequency. At best it will be similar in performance. But thats very unlikely.
Anyway, we will wait and see, but that information puts idea of Core M in different light...