The reluctant debut of the A14 processor
sixcolors.com
sixcolors.com
If true, I'm not really sure what to do about that. I doubt I can offload rust compilation to an iphone.
What applications benefit from this? The only use-case I've seen for AI in Apple's mobile devices is the face unlock.
Oh no. You could have double the CPU core and it wouldn't make 95% of common use case any faster. But NPU ( Neural Processing Engine ) and GPGPU have lots of potentials. Siri being the first one, When you take photos its uses NPU to adjust the image. Your Speech recognition for lots of different languages. Handwriting, Your Photo Gallery Face Recognition, on iPad Handwriting / Image to Text Recognition. And lots of thing likely to do with AR. And all of them has great impact to user experience previously not available.
Not to mention both GPU and NPU scales very well with transistor budget.
So the A14 is basically am A13 5nm Die Shrink with better GPU and Double the NPU. What I am interested in is the Die Size. Which I am expecting to be sub 80mm2.
I really love new tech and especially microprocessors but have trouble seeing how a phone breaks out of it's limiting form factor in any meaningful way right now that would make use of yet more processing power.
Since then we’ve got on-device powered AI voice assistants, Face ID, neural engine powered photo enhancement, health monitoring, custom silicon enhanced environment mapping AR. Goodness knows what else is coming.
The advent of custom tuned accelerators such as neural engines in silicon is opening up a huge new field of applications and capabilities. That’s especially true when they are linked to these new arrays of environment and bio-sensors. The X series iPhones are doing a list of things simply impossible on previous devices.
You may not care about some or all of these things, that’s fine they’re not all for everyone, but each one of them is important to some people.
I think what Guthur may reflect upon is that we got tons of new phone models pushed into the market every single year that do very little new compared to previous models.
So what if we get 50% more performance. Sounds impressive, but then you realize it is only used to create smiley animations, and other frivolous teenager nonsense.
I don't think I have really seen a point in phone upgrades since iPhone 4 or so. That is about 9 years ago.
Same deal with iPads, not much new anyone needs. There are a couple of things though which I do think has mattered. Retina display and the Apple pencil.
But honestly the priorities are quite different from what I would want. I love my Apple products but I really wished they where a bit more modular or serviceable. By that I mean that e.g. replacing the battery should be much easier. Memory should have been easier to expand or replace.
It's not just the apps that are missing, and the low resolution; I preferred its theme, and the smaller screen but it's just not fun to use anymore now that I've gotten used to the snappy everything, the dictation while driving, the better camera, etc. When I got the 6S, it was for app support - I didn't feel anything was missing - but I won't switch back to 4S even if it did get a security update. I'll likely get a new iphone when iOS 15 comes out and the 6S isn't supported; There's nothing I know I need it for, but I suspect I won't be satisfied with the 6S after I do have it.
Re:ipads, though - my childern's 2013 ipad mini is essentially as good as their new one for all their uses, it's only missing software and security updated that prompted getting a new one.
The apps stop working/updating after a while. So if you don't have a reason Apple gives you one: planned obsolescence
Because they do it for obvious reasons(i.e money)
>> when they support their devices sometimes 3x (5-6 years vs 2) what their competition (Android) is doing?
I guess they set a sensible lifespan considering the market they target(i.e high end)
Either way the result is the same: after a while you end up with a useless brick instead of an old/use-able computer.
It's also worth to note that at least on Android you can still side-load/install apps outside the play store while on Apple your device becomes totally useless/museum brick.
The number of competitors has little to do with planned obsolescence but
" Planned obsolescence tends to work best when a producer has at least an oligopoly "
Planned obsolescence is not related to advancements in technologies. Would you classify the Commodore 64 as having been designed with planned obsolescence in mind considering it ran on a 6502 when the competition at the time was transitioning to 8086? Of course not. So why is Apple deciding to put more advanced technologies into their devices planned obsolescence?
If you’re referencing the App Store, though, I can see where you’re coming from. But Apple releasing new technologies and not putting them on devices that can’t handle them (due to computing power) and developers using them (which prevents them from running on said old devices) is not planned obsolescence either. The developer chose to make it incompatible with the older devices, not Apple.
To answer your question:
today I can still deploy code on Commodore 64 but not on my old iPhone. How is that possible?
https://developer.apple.com/news/?id=03042020b
>> Starting April 30, 2020, all iPhone apps submitted to the App Store must be built with the iOS 13 SDK or later.
There's no law requiring vendors to provide features to install apps on devices with CPUs at all.
You probably own a dozen devices containing CPUs, running software, for which there is no practical way for you to install anything. Apple provides a certain set of features in their devices, if you don't think that's a good deal, there are other vendors you can go to.
Planned obsolescence my arse.
Things like Facebook and WhatsApp apps depend on the developer to keep the app compatible with their network API. I would expect mail to still work, depending on which service you use. Any custom hooks into GMail for example might fail if Google changes the back end interface.
The apps themselves should still ‘work’ as in run on the device just fine as they were built for that OS, it’s compatibility with external services that may have changed their APIs that is the main issue.
Submissions with the old APIs are no longer accepted so tell me, isn't that planned obosolence?
Basically I as a developer and owner of the device cannot update my own app, on my old device anymore because the OS does not support the new APIs and Apple does not allow me to submit updates with the old APIs or to upgrade the OS.
Your only choice is to not use Apple APIs that are changed often. In practice you can do that only on certain games or niche apps.
https://developer.apple.com/news/?id=03042020b
>> Starting April 30, 2020, all iPhone apps submitted to the App Store must be built with the iOS 13 SDK or later.
I've never used an iPhone, but Galaxy phones used to have a really stupid and annoying button for their intrusive personal assistant. Now they don't have any such intrusive hardware or software designs.
The newest versions of Android, which are only supported on newer devices, allow much more fine-grained security permissions. In particular, I like that they allow blocking background access to a capability if the app isn't currently in use.
The new AR camera apps are really fun.
WiFi 6 and 5G provide noticeable improvement in internet speed.
The higher definition screens without bezels provide better video experiences for the same device dimensions.
Allowing fingerprint reading through the screen makes for a better experience.
Wireless charging, particularly in the car with a magnetic mount, is very slick and you never will want to go back.
If you need a cable, USB-C is clearly better than micro-USB.
All of this stuff only became common in the past couple years. If your phone is 3 years old, you probably don't have any of it. You will be fumbling with cables on a much less capable device.
I can't imagine what other sorts of polish they will figure out, but betting against it seems like a historical loser.
I honestly think Apple could kill off all but professional photography and film with their own versions of photography kit such as Moment lenses, DJI Gimbals and associated AI correction tech to get rid of the shortcomings. Add in more powerful noise reduction in low light that their AI will soon enough bring and your phone camera is a better photographer than you will be without a great camera, training and practice.
The last real innovation in Intel chips in laptops and desktops I can remember that had a significant impact was the accelerated encryption and that’s years ago.
I’m honestly excited by Apple’s new chips going into laptops - Will we get webcam calls that have their image quality improved and background noise filtered by AI for example?
https://www.wired.com/2015/10/light-l16-camera/
Processing all that data, especially in real time, will use an increasing amount of processing power.
You'll have a DSLR level of quality and control, a depth map of every photo and video, much better post-shot refocussing, etc.
The iPhone already does this, outside portrait mode. It’s called Deep Fusion and triggers on well lit photos shot using either the 1x or 2x lens (aka not on the 0.5x, super wide, lens).
My 2018 iPad pro is sort of the same, but as I find myself using it for more and more things, I can see places where more speed would help. It's my goto device now for using LR and sorting and editing pics from my DSLR. It's already amazingly fast, but in cases like this more speed is still generally better.
Another interesting possible future is the desktop replacement, dock your phone into a hub that exports video and USB signals to connect a keyboard mouse and screen and do away completely with laptops for user's whose computing needs are modest. Samsung does this already with DEX but it's too limited for me to view as more than a cool prototype but with Apple's app ecosystem this could be quite useful.
In other words... the thing phones can continue to do for general processing power is dictate efficiency by necessity, and make that efficiency generalizable if there’s an interest in that investment.
> two high performance A13 cores are 20% faster with 30% lower power consumption than the Apple A12's, and the four high efficiency cores are 20% faster with 40% lower power consumption than the A12's. A13's eight-core Neural Engine dedicated neural network hardware is 20% faster and consumes 15% lower power than the A12's
> A14 up to 40% faster than A12, with 30% faster graphics performance than A12, and machine-learning performance is up to 10 times faster.
The thing that I found interesting in the A14 is that this is the first time I can remember a chip designer using a die shrink mainly to cut power usage instead of increase performance.
However after Anandtech's recent deep dive into Tiger Lake performance testing, you can see where cutting power use in the A14 cores might be a winning move.
>Here we present the 15W vs 28W configuration figures for the single-threaded workloads, which do see a jump in performance by going to the higher TDP configuration, meaning [Tiger Lake] is thermally constrained at 15W even in ST workloads.
Comparing it against Apple’s A13, things aren’t looking so rosy as the Intel CPU barely outmatches it even though it uses several times more power, which doesn’t bode well for Intel once Apple releases its “Apple Silicon” Macbooks.
https://www.anandtech.com/show/16084/intel-tiger-lake-review...
This has been a thing with RF Chip design for the 30 years I've been paying attention. Transceiver I used 20 years ago consumed 220 mW in receive mode. One I used 10 years ago, 50 mW. Current one uses 10mW.
I can't come up with a single other example where cutting power was prioritized.
Things have tended to go the other way with die shrinks that favor performance and often with ridiculous amounts of overclocking piled on top of that.
>Brian and I have long been hinting at the sort of ridiculous frequency/voltage combinations mobile SoC vendors have been shipping at for nothing more than marketing purposes. I remember ARM telling me the ideal target for a Cortex A15 core in a smartphone was 1.2GHz. Samsung’s Exynos 5410 stuck four Cortex A15s in a phone with a max clock of 1.6GHz. The 5420 increases that to 1.7GHz. The problem with frequency scaling alone is that it typically comes at the price of higher voltage. There’s a quadratic relationship between voltage and power consumption, so it’s quite possibly one of the worst ways to get more performance.
Also they reserved all of TSMC's 5nm capacity for the A14. Like that was a last ditch effort to make it _look_ faster after all?
> So either they put all of their engineering capacity towards a blistering fast desktop / laptop CPU
TSMC promises up to a 30% improvement in power consumption or up to 15% more performance over the current N7 node in the 5nm process. This relatively matches with Apple's claim this gen. So Apple might have put the engineering effort on desktops & laptops this time.
This is also true for Qualcomm so I’d expect to see similar, mainly process node driven changes in their new chips as well.