Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Tekguy0

macrumors 6502
Original poster
Jan 19, 2020
306
361
Do you think Intel's next line, Tiger Lake, will affect the transition to ARM for the Mac, and will Apple continue to wait or will they switch before then?
 
CPU development costs a lot of money and takes a lot of time. If Apple is indeed developing ARM for Mac. I can't see them doing all that work then waiting a year or more over a small x86 performance bump.
 
  • Like
Reactions: chabig and Tekguy0
Switching to ARM is still a rumour and just now one that doesn't make a lot of sense. Intel CPU's are still very much on a different level of complexity to ARM and I don't see Apple wanting to kill the Mac just yet unless they have something really incredible in the works.
 
Switching to ARM is still a rumour and just now one that doesn't make a lot of sense. Intel CPU's are still very much on a different level of complexity to ARM and I don't see Apple wanting to kill the Mac just yet unless they have something really incredible in the works.

My expectation would be that ARM would be a coprocessor. It's incredible energy efficient. This way OS tasks and built in software could be offloaded to the ARM chip. Stuff like powernap tasks, iCloud sync, notifications, mail, calendar, &c. It would also make it possible for the Mac to natively run the huge library of iOS apps. Leaving x86 for the more feature rich commercial and open source desktop software.

Having the Mac able to run iOS software could make it more palatable to many potential buyers. As there are a lot of useful apps with no desktop counterpart. Which I'm sure many people would want to run on their computers.
 
  • Like
Reactions: AlphaCentauri
Personally, I think an ARM-based Mac is wishful thinking from some Apple enthusiasts.

If ARM showed up on a Mac, it would be on the low-end MacBook -- not the desktop -- presumably a replacement for the discontinued fanless 12" MacBook 2017. But Apple has been pushing the iPad Pro.

Apple has made a strong and concerted push to expanding peripheral support on its iPad Pro product line. This is the most obvious clue of what Apple thinks the future direction is.

Another factor is completely out of Apple's control: the availability of high-quality desktop applications running on other ARM-based Windows. These would be the full-fledged Microsoft Office suite, the Adobe applications, etc. Unless there is a wide selection of those types of applications which could A.) be ported to run on macOS for ARM or B.) run in an emulator/virtual environment in Windows for ARM, I don't see the release of an ARM-based Mac as forthcoming.

Macs are less than 10% of the world's PCs. If an ARM-based Mac does emerge, it probably would represent less than 1% of the world's PC quarterly sales and even less of the installed user base. That is an awfully small number of units for a software developer to support.

If the user is mostly going to use e-mail, light office apps, web browsers, etc., there's no reason to do that on an ARM-based Mac.

Catalyst is a way for iOS developers to port their apps to macOS but there isn't much traction that I can see.

My guess is that there are ARM-based notebooks running macOS right now in a lab in Cupertino. They are probably sitting next to the AMD-based notebooks running macOS. In a nearby lab, there are probably ARM-based and AMD-based Mac minis running macOS.

An ARM-based Mac would have to give some sort of material benefit beyond an extra hour or two of battery life. There is nothing at the present time that would indicate that there is enough inertia for this.

Tiger Lake is not a factor. Apple is well aware of Intel's product roadmap as well as all of the delays, failures, and stumblings Intel has made over the years. In fact, Intel has given Apple some early access/exclusivity for some processors.

It's really a decision based on software availability and whether or not an ARM-based Mac would be a meaningful and useful departure from x86. Apple isn't going to do it to sell 100,000 units that can only do e-mail, a web browser and a handful of applications.
 
Last edited:
Do you think Intel's next line, Tiger Lake, will affect the transition to ARM for the Mac, and will Apple continue to wait or will they switch before then?

Tiger Lake will probably ship in 2020. It has already been leaked that Apple ARM isn't coming to 2021.


ARM for Mac is already probably siding behind Tiger Lake ship time ( at least on some system vendors. Apple may not be a 'first mover' there. ).

Tiger Lake is a subset of the Intel processor line up. It is only going to cover a the mid-upper range of the types of processors Apple uses in the MacBook Pro zone. It won't be a dramatic turn around for something like a "one port wonder" Macbook. As long as Apple is fixated on doing the latter then (some) Mac on ARM is likely still in motion for 2021.

What Apple is grossly missing is not cores to match up in some single thread benchmark drag race but robust I/O. Can't attach a discrete GPU to their ARM implementation. No Thunderbolt/USB4 . relatively weak USB breath . etc. The only Mac laptop they are really demonstrated readiness for is the "one port wonder" they dropped. ( and in part becaues they didn't have a 'good match' processor to go with their thinnest of thin objectives. )

Tiger Lake isn't the only option though. AMD has mid-upper range solutions that are respectively competitive also. They are 'missing' on robust vendor support on Thunderbolt but it is less of a broad I/O gap than the Apple solution has.

The question is how much does Apple want to go "split architecture" line up across the Mac ecosystem or not. There downsides to splitting a relatively small market into smaller segments. But there are also upsides of Apple chasing lighter/thinner of going with they are already making anyway for iPad Pro line up.
 
  • Like
Reactions: Tekguy0
Switching to ARM is still a rumour and just now one that doesn't make a lot of sense. Intel CPU's are still very much on a different level of complexity to ARM and I don't see Apple wanting to kill the Mac just yet unless they have something really incredible in the works.


ARM can be implemented in a range of CPU packages. The "instructure set" and "general architecutre" can be competitive.


The real question is why would Apple sink lots of money into a different ARM implementation track than what can be used across the iOS and iPadOS product lines. For iOS and iPadOS product Apple either uses the same processor in all models of that generation or uses a "hand me down" processors from some previous iOS iteration. ( At best, they have three implementation variations of 8-12 products. iPhone targeted one. a tweaked iPhone baseline bigger for iPad Pro , and a prune off 'big' processor one for the Apple Watch. And that is about it. ).

One reason why Apple's ARM implementation has outperformed the competition is because they have been largely monomaniacal in optimizing it for just one product ( latest gen iPhone). If they dilute their focus they'll start to be in the same boat as their competitors ( who don't put all their wood behind one arrow and even up with "pretty good" products on multiple fronts. )


Whereas the Mac product line has about has about 5-6 implementation variations over 6-7 products. And has get to the larger Mac products the volumes go substantially down. ( so for example the year run rate for the 'new' Watch processor is much higher than almost all of the individual Mac products. )

Apple is going to do it so that they can pass along lower prices to customers. LOL. ( https://www.macrumors.com/2020/05/30/13-inch-macbook-pro-ram-upgrade-doubled/ ). Appie has gotta spend a giant bucket of money on something. already did. https://www.macrumors.com/2019/12/02/apple-intel-modem-deal-completed/ ( don't be surprised when Apple's modems roll out that suddently putting a modem in some Macs becomes an insanely great idea. )

if Intel was the only x86 choice and they were continually screwing up then yeah... Apple would/will probably dump them. But they aren't. AMD has been executing increasing well for the last 3 years. If Apple tries to jump into the higher end desktop market it won't just be Intel they are competing with it will be AMD too. At the very top end it will also be other ARM vendors.


A "forked" ARM laptop shadowing the Windows "always on' with ARM play would be a reasonable move for Apple to make. But they don't have to sink tons of money to do that if just tweak the iPad Pro processor slightly.
 
Last edited:
  • Like
Reactions: Tekguy0
Do you think Intel's next line, Tiger Lake, will affect the transition to ARM for the Mac, and will Apple continue to wait or will they switch before then?

Tiger Lake in particular shouldn't have any effect. I think people attribute short-term / single-generation thinking much more than Apple does.

These:
  • Ryzen Mobile 4000 outperforming Comet Lake
  • Kaby Lake, Coffee Lake, Comet Lake, Whiskey Lake, Amber Lake having been mediocre bandaids on top of Skylake,
  • Ice Lake finally showing a bit of a light at the end of the tunnel
…I'm sure Apple has noticed all of them. And I'm sure they've factored a little into Apple's decision-making. But they probably care more about the long-term roadmap.

Is Intel having a hard time moving from 14nm to 10nm a fluke, or will they have similar trouble again moving to 7nm or 5nm? Is Ryzen just pretty good right now (and a fair amount ahead in Zen 2), or will AMD be consistently equal or ahead in the coming years? And then: is Apple's own ARM ahead on either of those?

(And, with AMD: can they deliver in the volumes Apple needs? Keep in mind no major laptop vendor defaults to AMD. Is volume part of the reason?)

Where Tiger Lake might matter is in terms of Apple delaying any (potential) transition to ARM by a year or two. But that's about it.
 
  • Like
Reactions: Tekguy0
My expectation would be that ARM would be a coprocessor. It's incredible energy efficient. This way OS tasks and built in software could be offloaded to the ARM chip. Stuff like powernap tasks, iCloud sync, notifications, mail, calendar, &c.

Intel has been toying with this without having to fork off a seperate set of binaries.


It would also make it possible for the Mac to natively run the huge library of iOS apps. Leaving x86 for the more feature rich commercial and open source desktop software.

Errr not exactly. The iOS apps presume a different OS Foundational library. Would be able to run them on something and a map the graphics output back to a sub window but that grosses ignores the fact that the iOS apps are going to be expecting a touch screen display to be there and there very probably won't be one. That isn't going to do much good.


There is likely going to be a "coprocessor" but it will likely be the same job allocation have now of a bunch of supportive tasks of security , optional much smaller, non-main touchscreen, and audio-video (and probably future AI/ML 'Siri smarts' ) work. Basically a follow on to the T2. More random users apps on the security processor is an oxymoron. Whole reason for being a discrete processor was to get away from the random user injected code.

Having the Mac able to run iOS software could make it more palatable to many potential buyers. As there are a lot of useful apps with no desktop counterpart. Which I'm sure many people would want to run on their computers.

that is what Catalyst and SwiftUI are for. Also Apple's incremental moves to a unified App store. The developers can do the ports to get around the user interface differences / gulfs between the products. ( primarily touch and smaller screen versus mouse and relatively much lager screens. ). Running a large number of Android phone apps with phone screen restrictions) really isn't helping Android tablets. ( not did it really help the iPad much either. That's why the push to get larger fraction of developers to do iPad specific ones. )
 
  • Like
Reactions: Tekguy0
Catalyst is a way for iOS developers to port their apps to macOS but there isn't much traction that I can see.

One problem for Catalyst is that Apple introduced two "path to porting and the future" mechanisms at the same time. A contributing factor is "which one" (Catalyst or SwiftUI) . It isn't quite the same but in the same ballpark as Carbon versus Cocoa. Only there was much more natural inertia behind Carbon.

The other substantive problem with Catalyst is that it has been substantively oversold. It is more a framework to make porting to macOS easier. It isn't the "just click the check box for macOS and your substantive app with automagically ports itself. ". There are variances of GUI and layout have to code for if want an app that adapts to the focus of each platform ( touch versus pointer , screen size breadth , etc. )

Catalyst is a good match for developer groups who were already in the business of commuting to porting apps to both app stores. But a huge flotsam of iOS apps just out to get folks to fire it up so can present ads and track data to make money.... macOS ports probably don't have much of a pay off.






My guess is that there are ARM-based notebooks running macOS right now in a lab in Cupertino. They are probably sitting next to the AMD-based notebooks running macOS. In a nearby lab, there are probably ARM-based and AMD-based Mac minis running macOS.

ARM based mini ? Probably closer to a ARM based AppleTV running MacOS. They already have that hardware for a baseline. Don't need to create anything.

AMD mini ? Should ... that would make some sense.
Since some AMD codenames for AMD processors show up in macOS betas ... there probably are at least "bake off" design competition test harnesses present inside of Apple.

[ if only to jab at Intel for price reductions. Apple likes having two (or more) suppliers so can pit them against each other for lower component/services prices. The notion that Apple wants no choices is a kind of goofy. ]


An ARM-based Mac would have to give some sort of material benefit beyond an extra hour or two of battery life. There is nothing at the present time that would indicate that there is enough inertia for this.

the latest MBA thermally throttling ( admittedly with a dubious fan cooling implementation) and the butterfly keyboard disaster ( spawned out of their thinnest of thin MacBook objective) it isn't so much battery life but being able to get about the same battery life in a smaller/thinner/lighter enclosure.


Apple isn't going to do it to sell 100,000 units that can only do e-mail, a web browser and a handful of applications.

Their ARM implementation may have problems with 8-16+ GB RAM workloads and up but there is a probably a reasonable amount of stuff it can do. The iPad Pro is no slouch with a !2Z and a 14Z is probably more than respectable up against the 8GB 8th-gen MBP 13" they are selling now. I/O ports/expansion , RAM capacity and storage capacity is where there is a probably a bigger gap than in the kinds of apps can run.
 
  • Like
Reactions: Tekguy0
One reason why Apple's ARM implementation has outperformed the competition is because they have been largely monomaniacal in optimizing it for just one product ( latest gen iPhone). If they dilute their focus they'll start to be in the same boat as their competitors ( who don't put all their wood behind one arrow and even up with "pretty good" products on multiple fronts. )

I see no basis for these assumptions.

Sure, the main use case of Apple Ax is the iPhone. But there are few features in it that are all that iPhone-specific. You could count the Neural Engine, maybe, but for one, even that has applications on the Mac as well, and two, they can simply leave that coprocessor out if they like — just like the T2 probably lacks the "M10" motion coprocessor, despite being A10-derived.

Beyond that, it's not that phone-specific. It's simply an ARM CPU that heavily favors single-thread performance. This bet, which contrasts with Qualcomm going to more cores earlier than Apple did, works for multiple reasons: for one, Apple's applications typically don't have a garbage collector thread running. So that's a big chunk of background thread work (not to mention RAM use) gone right there. And two, a ton of work these days is still mostly single-threaded; JavaScript, for example.
 
  • Like
Reactions: Tekguy0
I know little but there seem to be too many strong rumors for me to not think that Apple is moving towards ARM based macs. But I think it will be a relatively slow migration. I suspect the initial product will be a relatively low end Macbook variant that's kind of an iPad Pro in a Macbook platform. Probably not too unlike a current iPad Pro with one of the new keyboards, and track pads. And of course we now have the ability to use a mouse. There's already a lot of software that functions quite well for doing most of the things most people do most of the time. For me, the iPad suffers as a data generation tool. I've not tried the new keyboards, but I hate trying to type much on an iPad. Give me an iPad Pro with built in keyboard, touch pad, ability to use a mouse, ability to connect to external memory devices and it would be a terrifically useful device. Sort of a Macbook Air with iPad pro internals. I'd expect it will take quite a long period of time before there would be a complete transition. Initially it's not going to meet the needs of folks using specialized software or who are doing highly demanding tasks. So I think this will have very minimal or no effect on the next generation Intel processors. Now, a move to AMD processors...
 
I see no basis for these assumptions.

Not really an assumption when Apple says that is what they are doing.

"... Srouji’s team found itself interacting regularly with other departments, from software programmers, who wanted chip support for new features, to Ive’s industrial designers, who wanted help making the phones flatter and sleeker. An engineer who sat in on Srouji’s meetings remembers senior managers preparing extensively for presentations, because his support was critical for getting new features approved ... "

The industrial design team digs "holes" that the silicon team is suppose to get the product out of. That is why Apple's broader use of the processors design for getting out of those particular product "holes" is typically as a "hand me down" processor that is already 'paid for' in terms of R&D and just has more than enough "horsepower" for that broader product placement.


" ... “We've got our own custom-designed performance controller that lets you use all eight at the same time,” Shimpi told Ars. “And so when you're running these heavily-threaded workloads, things that you might find in pro workflows and pro applications, that's where you see the up to 90 percent improvement over A10X. ..."

Those tuned threading loads are iOS and several selected iOS apps or random Linux and Android apps ? For example,

"... On iOS, 429.mcf was a problem case as the kernel memory allocator generally refuses to allocate the single large 1.8GB chunk that the program requires (even on the new 4GB iPhones). I’ve modified the benchmark to use only half the amount of arcs, thus roughly reducing the memory footprint to ~1GB ..."

So how much time did the chip designers spend on MMU performance for 2-3GB user level memory workload and memory pressure which the OS blocks. ?

Sure, the main use case of Apple Ax is the iPhone. But there are few features in it that are all that iPhone-specific.

it is just as much an issue of missing features as it is ones that are present. Even those present are tuned to which workloads in the design phase.

This bet, which contrasts with Qualcomm going to more cores earlier than Apple did, works for multiple reasons: for one, Apple's applications typically don't have a garbage collector thread running. So that's a big chunk of background thread work (not to mention RAM use) gone right there. And two, a ton of work these days is still mostly single-threaded; JavaScript, for example.

this is more so about Qualcomm spending time and resources on a "server" processor which they eventually walked away from.

That was part of the whole issue of how Apple got to 64-bit ARM before everyone else did. Almost all of the other ARM implementors viewed 64 bit ARM as how to break into the higher margin server processor market. So had different timelines assigned to that implementation path. Apple didn't really care anything about the memory address expansion ( iOS proactively kneecaps apps in memory allocation as soon above). 64 was a path to dump some legacy ARM instruction baggage.. so Apple took that early. That was in part driven by the fact that iOS and their app software didn't want to use that other legacy stuff either and it lowers the workload on compiler/tools for Apple.... so they jumped on that path quicker.

Baseline ARM is trying to do multiple things. Apple ignores significant parts of those. The other implementors are often in business of building implementations for other folks to use to build things with. Qualcomm rolls out more Snapdragon processor variants in a year than Apple drops iPad+iPhone models in the same year.
 
  • Like
Reactions: Tekguy0
Not really an assumption when Apple says that is what they are doing.

"... Srouji’s team found itself interacting regularly with other departments, from software programmers, who wanted chip support for new features, to Ive’s industrial designers, who wanted help making the phones flatter and sleeker. An engineer who sat in on Srouji’s meetings remembers senior managers preparing extensively for presentations, because his support was critical for getting new features approved ... "

The industrial design team digs "holes" that the silicon team is suppose to get the product out of. That is why Apple's broader use of the processors design for getting out of those particular product "holes" is typically as a "hand me down" processor that is already 'paid for' in terms of R&D and just has more than enough "horsepower" for that broader product placement.

Oh, sure, but at the end of the day, again, it's mostly an ARM-based CPU design that favors single-threaded performance. It's not a particularly poor fit for most Macs.

It might be a poor fit for the many-core high-end Macs like the iMac Pro and Mac Pro, but I don't think that's a priority for moving away from x86 anyway; the Xeon W is fine.

it is just as much an issue of missing features as it is ones that are present. Even those present are tuned to which workloads in the design phase.

OK, but which features are missing? If anything, Fusion (similar to big.LITTLE) would be a plus on MacBooks that Intel can't currently offer, and seems to only plan to start offering by Alder Lake a year or two from now.

I guess it's missing something like Turbo Boost, but I don't think that's critical on the lower end.

Baseline ARM is trying to do multiple things. Apple ignores significant parts of those. The other implementors are often in business of building implementations for other folks to use to build things with. Qualcomm rolls out more Snapdragon processor variants in a year than Apple drops iPad+iPhone models in the same year.

I mean, most ARM design licensees walked away altogether. Whether it's Qualcomm Kryo, Samsung Exynos or Nvidia Tegra: they used to have their own custom designs, but have mostly moved towards Cortex-derived designs. Apple, in contrast, has moved the opposite direction: up to the A5X, they used Cortex-derived designs, but since the A6, they've moved exclusively to their own custom design.

So if Qualcomm wants to whine that they didn't hit 64-bit first: well, maybe they should've tried harder?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.