Hacker News new | ask | show | jobs
by mmastrac 3424 days ago
I think that in the long run you'll see a hybrid ARM/x64 Mac sharing the same case in the same way we have hybrid graphics.

If you look at most of the processes running on a Mac, how many of those need to be running on the power-hungrier Intel chip? You could easily offload much of the kernel, Dock, compositor, Twitter desktop, etc to a lower-power ARM. In the case where you aren't running any x64 apps, you could then power that chip down and run solely on ARM.

8 comments

This sounds incredibly kludgey and at odds with Apple's de-investment in macOS. (As in "reduced investment", not "divestment".) Adding massive complexity to eke out another hour or two of battery life seems like a poor choice when there are so many things to be fixed in the OS.
Why so? A Mac is a very high ASP item - why not, for example allow iOS apps to run in macOS (using the ARM processor).

Ultimately, the 2016 MBPs are a great example of why Apple needs to divest from Intel. The poor timing for Kaby Lake mobile processors and lack of low-power RAM were both Apple's reasons stated when the topic of "the MBP isn't keeping up with the competition" arises.

And who's to say that macOS.next will be anything similar to the current macOS?

Apple wouldn't need an ARM chip to allow iOS apps to run on Macs. I would think that all iOS apps are already compatible with x86. When a developer writes apps today using Xcode and runs the app in the simulator, they are running an x86 compiled version of their app linked against an x86 version of the iOS framework. It wouldn't be that much of a technical hurdle to polish the experience and officially support fat binaries that run on x86 and ARM.
Only in theory. In practice the existing base of software in the App Store would cause compatibility issues.

Intel based Android devices have an ARM emulator on board (libhoudini) and claim to be ARM when querying the store, in order to not have a software offering disadvantage.

I'm not saying it's a good idea, but we are talking about Apple allowing IOS apps run natively on Macs. iOS apps are already running natively on macOS - but only if you have Xcode. Apple could really just include the x86 based iOS framework that it already has in the next version of macOS and tell developers if you want to run on Macs, just modify your UI to support a third target - you are already probably targeting iPhone and iPad - let Xcode bundle the x86 build its already doing while you're testing.

As far as Intel based phones, that's even easier, tell developers they are realeasing an x86 based phone in the next year, either you bundle an x86 version - again you're already testing an x86 build every time you run the emulator - or you'll lose compatibility. Apple has never been afraid to abandon apps when it transistioned processors.

The Android situation is different, if people have a choice between buying an Android device that is 70 percent compatible and one that is 100% compatible. The one that is 70 percent compatible is at a disadvantage. But if people want an iOS device and Apple switches to Intel and they lose some apps what choice do they have?

I'd be amazed if Apple ever gave up ARM for x86. Has Intel improved that much in comparison to ARM for power efficiency?
why not, for example allow iOS apps to run in macOS

That's not what the article is talking about and it's a bad idea for UX reasons (no touchscreen).

And who's to say that macOS.next will be anything similar to the current macOS?

Don't even go there; a vocal minority of Mac users including me do not want anything like that. And getting back to the article, if the OS is radically different then presumably it doesn't need both Intel and ARM processors.

> That's not what the article is talking about and it's a bad idea for UX reasons (no touchscreen).

Windows 10 Apps seem to have bridged the divide between mobile and desktop, and run fine on both.

It wasn't instant, Windows has been trying to bring this divide closer since Windows 8... and after so long they're pretty much succeeded with the Surface tablet/laptop.

> Don't even go there; a vocal minority of Mac users including me do not want anything like that.

I'm a Mac user too. Unfortunately, Apple is notorious for doing what they think is right, regardless of their user base. Need I say "lightning headphones" or "no magsafe, just USB-C"?

Both of these strike me as positioning for the future Apple wants despite the concerns the vocal parts of their userbase.

Skate where the puck will be, not where it is.
No reason they couldn't put a touch screen on a MacBook or imac, though I guess they can't guarantee one on a mac pro
Well, one reason is that touch screens for desktop OSes suck. Holding your arm out screen operations for any length of time is awful, and wiping the prints off your screen on your pant leg doesn't work very well on a laptop.

That's not to say Apple doesn't release terrible things at times (hockey puck mouse[1]?), but I have trouble imagining Apple releasing one of those. (Laplet? tabletop?)

[1] I actually think that may have been calculated - it is impossible that Jobs didn't know it was a shitty mouse, but alongside the iMac, it got a ton of attention at the time Apple really needed it.

> Well, one reason is that touch screens for desktop OSes suck

They work perfectly well on desktops like the Surface Studio, where you can bring the screen down to a drawing board angle. They also work well with a "desktop OS" on convertible laptops and 2-in-1s with multiple use modes including "tent" and "tablet".

They're actually not too bad on real desktops with vertical screens, especially if you're standing up. (Try putting an all-in-one in your kitchen.)

It would be foolish to assume that just because you have a touch screen that you have to touch it all the time. You don't. You can still use a mouse, a pen, an air-mouse, a games controller, and several other things. Having a touch screen doesn't make you do anything you don't want to do, it just gives you an extra option.

It's a bit like claiming that if you use a mouse you can't use keyboard shortcuts. Really, that's not how it works....

I have a hp spectre with a touch screen and windows 10, and it certainly doesn't suck. Great for sticking in tablet mode and browsing, organising photos etc.
No touch screen at the moment. Apple could be looking at making a hybrid ipad/mac.
That would cannibalise sales of the iPad. Its why they have yet to release a proper iPad Pro, just a gigantic iPod touch with a stylus.
Macs cost more so why would that be a bad thing?

iPad sales were down by 22% in the last quarter and are roughly half what they were at the peak. As someone has pointed out, the iPad sales chart is looking like the iPod's, except it might be going down a little faster...

http://www.zdnet.com/article/the-ipad-is-on-life-support/

Tim Cook's "why on earth would you buy a PC?" routine isn't looking quite as hot now iPads are the fourth ranked revenue earner at Apple, behind the iPhone, Macs, and Services.

>The poor timing for Kaby Lake mobile processors [was one of] Apple's reasons stated when the topic of "the MBP isn't keeping up with the competition" arises.

I've heard similar sentiments quite a few times. However, there's little reason to believe that Apple would have been able to meet their needs better than Intel could have. It's not like Intel's development schedule was misaligned with Apple. They just couldn't get the processor out in time.

> why not, for example allow iOS apps to run in macOS

Never say never, but that is the sort of thing Apple would never do.

I believe the Touch Bar on new MacBook Pros is a modularized component running a version of iOS on ARM.
It's running a stripped-down watchOS actually
Hadn't heard that, but I can believe it. It is also entirely different than users running Infinity Blade or whatever.
Is apple divesting its mac offering? Sounds like FUD.
This seems... suspect. There are definitely serious concerns (see the Mac Pro), but given the continued yearly releases of macOS and and the (much maligned, but a company spanning effort) touchbar, it hardly seems like they have plans to divest.
Divestment is an oversimplification of the direction Apple seems to be taking the Mac. Consolidation would be a better term: trying to prune the product line into a lineup of hits and nothing else, often to the exclusion of non-mainstream (professional) users.

So the test to apply is: will this enable a headline feature? In this case, if it lets Apple put a significantly bigger battery life number on the product page, I'd bet on it.

I mean, I'm not a fan of the new MBP myself, but I don't think they're ignoring professional users. Didn't their MBP sales go UP?
"often to the exclusion of non-mainstream (professional) users"

The allusion is to the Mac Pro, not the Macbook Pro.

There was pent-up demand from people who had been holding off buying the old model....
Because you can't buy something else from the PC industry?

If there is pent up demand, it's because Apple has been treating their costumeira much better than every other OEM.

Proof?
Apple has done many, many things to eke out a couple more hours of battery life. And there has always been a market for those improvements. Same with making thinner cases. it doesn't seem like much, but it sells.
> Apple's de-investment in macOS

Citation needed.

"We love the Mac and are as committed to it, in both desktops and notebooks, as we ever have been."

http://www.independent.co.uk/life-style/gadgets-and-tech/fea...

Actions speak louder than words, and it's been 3 years and 1 month since the Mac Pro saw any action.
They upgrade macOS every year. New iMacs should be out in March or April. Sure, something is wrong with the Mac Pro, but the overall Mac investment is still strong, perhaps a little slow at the moment.
I think Ben Thompson had the best take on this. Something was fucked up bad in the product, and they aren't budgeted to make the necessary capital investment so soon to dig out a marginal product that nobody buys anyway.
> [...] to dig out a marginal product that nobody buys anyway

I wonder if they could expand the Pro desktop to be for serious gamers as well as dev's, videographers, animators etc. It'd be interesting to see what Razor and Alienware make on their systems - could it be enough to justify a presence in the high-end desktop market?

As for bringing gamers to the platform, Apple already has experience with this...from the losing side at least. Originally Halo was meant for the Mac however Microsoft purchased Bungie in 2000 and much like GoldenEye for the N64, Halo was the game that sold enough Xboxes to make the platform solid. Apple would similarly need some top-tier titles to encourage people to embrace an expensive desktop.

Apple's Anobit acquisition seems to be what's giving it the edge in the NVMe arena. They have experience with quiet cooling solutions. The touch bar would be good for key-mapping in games. Back-lit keyboards are their jam. They have a serious investment in Imagination Technologies Group who are behind the PowerVR architectures. Smaller, more reliable PSUs. Their austere industrial design is second to none and deeply contrasting to the garish Razor and Alienware designs.

It would come down to whether there's money to be made in the area, and I suspect most serious gamers build their own systems.

> It would come down to whether there's money to be made in the area, and I suspect most serious gamers build their own systems.

Spot on. The ultra-performance professionals and the big-spending gamers often do their own builds and then keep upgrading them.

Apple has always seen itself more of a sealed-box consumer appliance supplier, basically de-skilling computing.

But the comment was about MacOS not the Mac Pro
I've read that there is no longer a separate macOS development team; it's been folded in to the iOS team. I think that's a very bad idea.
I don't think we'll see an Arm/x86 hybrid, that wouldn't make sense as they're completely different ISAs and they either need to cross translate between them, or dobble compile everything.

But a heterogenous CPU certainly does make sense. A super efficient, but not particular fast core for specific tasks that are more or less always running, and the beefy full core for regular work. If it eeks out another 2 hours of battery life, I can see Apple doing it when they're redoing a CPU from the ground up.

Too late, Apple already invented that: https://en.wikipedia.org/wiki/Apple_A10

The A10 is the new chip in the iPhone 7. It has two fast cores and two energy-efficient cores.

EDIT: Sorry, "invented" was the wrong word. Apple is already doing that. My point was that the parent's idea isn't new.

Apple already invented that

ARM already offered this (A7+A15|A17, A53+A57|A72|A73), and Android phones have been using the design for years.

That is just an argument for them doing it for their laptop chips as well. It makes perfect sense from a battery life point of view.
Actually, eking out another hour or two of battery life at great difficulty sounds exactly like the kind of thing Apple, but few others, would do, if you think about it.
On the other hand, they're commoditizing Intel further and differentiating from the competition.

HP and Dell have gotten smarter and are making better devices. Apple has powerful resources in the ARM space that their competition cannot match -- they are 100% dependent on Intel, who are having issues shipping product.

Either your prove that assertion of "reduced investment" or you retract your comment.

Apple haters...

I see this being the first step as a multi-year transition away from Intel, rather than some sort of merger of iOS and macOS or a sudden architecture switch like PowerPC to Intel.

For now, the ARM will be a low-power co-processor handling a few tasks via specially compiled extensions. However, as Apple continues to iterate on the AX processors, and they, along with developers, figure out how to move more stuff onto ARM, the role of the Intel chips will decrease until they are the low-power co-processor running only legacy stuff which can't be easily ported or emulated. Until one day they just go away.

It makes sense that they'd want to own their core processor tech – they've got the talent and in a few years time will have the technology. It seems that each architecture change they've made has been driven by a supplier not being able to deliver, from Motorola with 68k to Motorola (again) and IBM with PowerPC and now Intel.

I mostly agree with this, but I still haven't figured out what the end-game for this might be. I believe that they'd just emulate low-power-requirement x86 and x64 code rather than keeping a low-end x64 chip in the laptop.

If a significant number of MBP customers are running VMWare or Parallels, however, this might seriously hurt the timeline for moving to 100% ARM. In that case they'd either have higher-end Macs with the x64 chips (potentially complicating the product line) or work on adding x64 emulation acceleration to their Ax chip line.

There already are ARM chips in the new MBPRs, for example - the touchbar and touch ID with its secure enclave runs on ARM. I could easily foresee them putting in more powerful ARM chips and offloading some tasks to them.
Hope it's less crashy/buggy in the future than my current touchbar :/
Funny, because often my touchbar works, but neither my trackpad or keyboard work.
Why doesn't Intel include an Atom core on their CPUs as a next level speed step, like ARMs big.little? It has to be much easier to keep both the high and low power parts the same architecture.
That would probably be less power efficient than just running a single core alone at reduced clock.
Which is why the Atom was discontinued, IIRC: the mainline chips power efficiency had been reduced enough to make the Atom much less viable.
Atom hasn't been discontinued; low-end "Celeron" and "Pentium" chips are actually Atoms.
Both Pentium and Celeron are used for standard mainstream dual-core chips and for Atom chips (eg. J4205, Celeron G3900, Pentium G4400)... so in a way they are Intel's trash-can brand.
I think you meant the mainline's efficiency increased. :)
The Playstation 4 does this as the south bridge is really just an ARM chip. So I think you are right.
And the PS3 before it. And the WiiU, and the Wii. And every AMD chip since about 2013 (AMD Platform Security Processor, their equivalent of Intel's ME).

Honestly, pretty much every complex electronic device (and quite a few not so!) out there has a bunch of ARMs inside of it doing system management. This is borderline a non-story IMO.

Hell, an ARM7TDMI was a pretty common core to use on SD cards a few years back. The damn things are everywhere.

> ARM7TDMI was a pretty common core to use on SD cards

Which is crazy, because that was the main CPU in the GameBoy Advance.

The PS4 southbridge has 256MB of RAM, runs FreeBSD, and downloads game patches and such. This seems far more ambitious than a BMC/ME.
Anyone interested in this should watch the CCC talk: https://www.youtube.com/watch?v=-AoHGJ1g9aM
Direct video link: https://media.ccc.de/v/33c3-7946-console_hacking_2016

The uploader seems to have recorded themselves playing the video and it's missing the start and has buffering issues,

Yeah dude that talk was awesome.
Not really, the ARM in the SB is much more like a ILOM system; not much different from the various processors scattered around hardware these days, be it ARM or some other more niche architecture, or even special-purpose archs.
They'll probably follow what they just did with the A10, at least in the near term. It has a low power mode and a high power mode, but doesn't run them simultaneously. Too complicated. It actually has 4 cores (2 low, 2 high) but can only run one pair at a time.[1]

So while that is similar on its face to hybrid graphics, I'm not sure it would be feasible to offload background tasks to the low power mode as it would be constantly switching modes every few milliseconds. More likely this architecture is better suited to Power Nap where the whole device enters a low-power state with restricted functions. That also sidesteps the issue of having to compile every third party app for two architectures. They can just implement the Power Nap stuff for ARM.

[1] https://en.wikipedia.org/wiki/Apple_A10

Doesn't that complicate the mainboard design vastly? Memory & caching, clocks & time, concurrency, etc.
Both CPUs are connected with a PCI bus and as far as the main SOC is concerned its a bunch of peripherals on PCI

https://media.ccc.de/v/33c3-7946-console_hacking_2016

The CPUs run different OSes and there is no ability to move tasks between them; this is rather unlike a hybrid or "big.LITTLE" approach.
Thankfully Intel has already done alot of new work in this area - allowing PCI devices (GPU, Xeon Phi) etc, to share computing resources more evenly.
Definitely going to complicate things but it could easily end up worth it for a laptop if it can reduce power usage to almost that of a phone when not doing anything particularly demanding.
I know it's not really what you mean, but isn't the touchbar powered by an ARM chip already and also the secure enclave?
Yes, it runs on a "T1" SoC, but the software is a watchOS derivative.