Apple to start making its own Mac ARM chips in 2021 instead of using Intel. What say ye?

RavinDJ

Supreme [H]ardness
Joined
Apr 9, 2002
Messages
4,440
I'm curious to see what people are thinking about this concept. Apparently, their in-house chips will have double the cores and, consequently, double the threads of the Intel chips. But they didn't say anything about the speeds.

In any case, are Intel and AMD SOL due to Apple's decision? Thoughts???

[edit]
As of 04/23/2020, I am getting closer and closer to the infamous 4,096 posts 😁
[/edit]
 
They won't use them across the board for everything.

So long my shit runs, I don't care.
 
They won't use them across the board for everything.

So long my shit runs, I don't care.
I think if they perform well they will ditch intel entirely in a shorter timespan than you might think.
Existing applications will be emulated and a timeline will be issued for all applications to move over just as they did with the transition from PowerPC to x86.
 
I'm hesitant to leap to ARM for a while since there are likely apps I'll need to keep running that won't get the necessary binaries, but I can see it taking off if Apple really knows what it's doing.

Remember, when Apple introduced the A4 people wondered why it had bothered with its own chips. Within a few generations it was beating other chip designers on a regular basis (look at how Qualcomm was caught off-guard by the leap to 64-bit). It'd be funny if by, say, 2025 you had to buy a MacBook to get the fastest laptop on the market.
 
It is interesting, I wonder if they plan to make things run through emulation/how well it will run. Also, how well it will perform vs. the competition using the same/similar programs.

I am looking to upgrade my 2013 MBP, but not sure if I am willing to place my trust in a new chip just yet.
 
The Mac Pro, is Xeon, and barely sells. The macbooks, sell good amounts and are Core iX's. Sell better.

Why the iphone can pull off ARM? Because it sells gazillions, everybody uses one. We STILL can't efficiently power a competitive smart phone with x86. lenovo tried with Intel. Didn't work out.

I don't see the MacPro going arm and being taken seriously. That means they'll have a 3 different architectures to deal with and they'll piss off a lot of software vendors. I don't see it ending well.
 
Switch the Mac to ARM and lose Boot Camp? I don't see it happening, given how being able to natively run Windows was one of the biggest reasons for the Intel x86 switch to begin with.

Yes, I know there's Windows 10 on ARM now, but if the Surface Pro X reviews were any indication, it's a compatibility jankfest compared to tried and true x86. Drivers don't exist for a lot of peripherals, games and other apps that use certain APIs tend to act up, stuff like that.

Then again, I suppose Apple users are no stranger to compatibility jankfests, given how often a new version of OS X/macOS will break something, most notably 10.5 Leopard dropping Classic Mode even for PowerPC G4/G5 users, 10.7 Lion dropping Rosetta for PowerPC binary support (which was a pretty smooth transition beforehand, like the 68k -> PowerPC switch), 10.14 Mojave sticking it to NVIDIA and requiring Metal-compatible GPUs for good performance, and 10.15 Catalina dropping 32-bit Intel binary support.

It's not that I don't think ARM will perform well enough in a desktop environment; everyone conveniently forgets that ARM, like PowerVR, was originally designed for desktop use and saw far more success in the mobile/embedded market. It's just that the IBM-compatible legacy of software spanning literal decades for x86 is too enticing for most of the market to move on from, unless an x86 system can be emulated in a convenient sandbox/VM with maximum compatibility and performance.

Note that even the OS 9 Classic Mode used up through 10.4 Tiger was not quite fully compatible with certain programs, despite being a generally good implementation; it's because of this that MDD Power Mac G4s still command a premium over the newer G5s, since they can boot OS 9.2.2 natively. (And, yes, despite the lack of preemptive multitasking and memory protection, some people really do favor OS 9 for production work, mainly the DAW crowd.)

You can bet that if the Mac Pro 7,1 winds up being the last Intel Mac by some stretch, it'll hold its inflated value just for that reason alone.
 
Switch the Mac to ARM and lose Boot Camp? I don't see it happening, given how being able to natively run Windows was one of the biggest reasons for the Intel x86 switch to begin with.
This was a bulletpoint for sales, not a 'biggest reason'. The biggest reason for the switch to x86 was the poor performance and flexibility of the previous CPUs.

The biggest draw for Apple aside from their ecosystem is their hardware; in every category that they compete, they make the best or second best at any given time, and all their stuff is designed to work together.

Yes, I know there's Windows 10 on ARM now, but if the Surface Pro X reviews were any indication, it's a compatibility jankfest compared to tried and true x86. Drivers don't exist for a lot of peripherals, games and other apps that use certain APIs tend to act up, stuff like that.
That's Windows; this isn't an issue for Apple. They've actually done all of the hard parts, and done it far more successfully than Microsoft has, successful in large part due to the control that they exert over their ecosystem.

Porting over to 'ARM Macs' will be a small exercise for those companies that already do desktop and mobile releases for Apple hardware.

x86 is too enticing for most of the market to move on from
Well, Apple isn't most of the market. Further, Apple builds appliances, for all the good and bad that entails.

Reality is that new buyers will likely be enticed by products that do what they need them to do fast enough or faster while exhibiting significantly longer battery life. That's really what ARM promises.

[I'll point out that this isn't impossible for x86, but rather, that level of specialization in terms of hardware is simply not a market currently worth pursuing as the producer would have to edge out existing manufacturers for design wins. Intel was very close but quite likely saw the protracted fight to get a return on their efforts to outweigh the potential benefits.]
You can bet that if the Mac Pro 7,1 winds up being the last Intel Mac by some stretch, it'll hold its inflated value just for that reason alone.
Agreed on whichever is the 'last', without a doubt!
 
This was a bulletpoint for sales, not a 'biggest reason'. The biggest reason for the switch to x86 was the poor performance and flexibility of the previous CPUs.

The biggest draw for Apple aside from their ecosystem is their hardware; in every category that they compete, they make the best or second best at any given time, and all their stuff is designed to work together.
The G5 was indeed IBM's big NetBurst moment - too hot and inefficient. It's not uncommon to see Power Mac G5s basically warp their logic boards from the sheer heat and cause solder joints to fail, bricking the poor things, whereas the worst that tended to plague their G4 predecessors was bad PSUs (especially the MDDs).

I'm sure that the general inefficiencies of the G5 making them wholly unsuitable for new PowerBooks was a big reason behind the switch, too. Even back then, Apple was trying to push things in the "laptop first" direction they're clearly taking now, knowing that most people don't need a beefy desktop workstation, just something that's powerful enough while still being mobile, even if that something was the famed 12" PowerBook G4.

Then again, I'm not most people, and neither is anyone else on this forum...

That's Windows; this isn't an issue for Apple. They've actually done all of the hard parts, and done it far more successfully than Microsoft has, successful in large part due to the control that they exert over their ecosystem.

Porting over to 'ARM Macs' will be a small exercise for those companies that already do desktop and mobile releases for Apple hardware.
I figured the 68k -> PowerPC transition was easy because it was pretty much seamless from an end user standpoint (seriously, I have yet to find a 68k app or game that breaks when running on a Power Mac), and PowerPC -> Intel was almost as smooth thanks to Rosetta, though I still notice what seems to be a performance hit in PowerPC games running on my 2008 Mac Pro 3,1.

To me, Apple's moves like dropping Classic Mode, PowerPC support, and recently, 32-bit Intel binary support are their way of strong-arming developers into getting with the times and making their codebase maintainable for the future, at the cost of making things a headache for the end user in the short term, as well as crazy people like me who still fire up old games from the '90s on computers in the 2020s if they can get away with it.

Most Mac developers are all too eager to advertise the new APIs they're supporting, though; I saw that a lot back in the (Snow) Leopard days, Core Image this, Grand Central Dispatch that. They're usually more willing to embrace the architectural changes.

Well, Apple isn't most of the market. Further, Apple builds appliances, for all the good and bad that entails.

Reality is that new buyers will likely be enticed by products that do what they need them to do fast enough or faster while exhibiting significantly longer battery life. That's really what ARM promises.

[I'll point out that this isn't impossible for x86, but rather, that level of specialization in terms of hardware is simply not a market currently worth pursuing as the producer would have to edge out existing manufacturers for design wins. Intel was very close but quite likely saw the protracted fight to get a return on their efforts to outweigh the potential benefits.]
The Macintosh was indeed an appliance from its inception - plug and play, simple to use, but also closed off from any real upgrades (not that it stops all those SE/30 enthusiasts much).

Problem is, I've never been much of an appliance guy outside of the kitchen. If it's a computer, I want to tinker with and upgrade it. If it's a car, I'm definitely going to tinker with and upgrade it. Got too much of that enthusiast mindset running through me, where "good" simply isn't good enough when I know it can be better with a few simple tweaks.

You might notice that my fave Macs - the Power Macintosh 9600, the Mirrored Drive Doors Power Mac G4, and the classic G5-derived Mac Pro cheese graters - all follow the same template of an upgradable, expandable desktop workstation that you can easily tinker with by slapping new CPUs, graphics cards and drives in as desired. Especially GPUs, since Apple is notorious for falling behind on those.

iOS devices fill the ARM appliance role nicely as it is, though there's still a few irksome limitations I find with it, even with jailbreaking. Apple wants to keep those distinct from macOS, yet integrate the two via Continuity, Handoff, and more recently, Sidecar. It presumably helps with ensuring that each OS has a focused ecosystem.

I guess what I'm really thinking about is "What separates an ARM-based Mac from an iPad Pro in practice?"

Are they just gonna withhold Xcode to whatever gets to wear the Mac label, as usual, such that you still can't develop your own iPad apps on the iPad itself? Can you still use PCIe cards and other such expansion hardware on desktop ARM Macs? Will just being able to plug in USB peripherals and have them work without fuss still be a Mac thing rather than an iOS thing?

We already know the typical person who just wants an appliance for Web browsing would be fine with an iPad and the new trackpad-laden keyboard, so there must be some reason they still favor using MacBooks instead.
 
Back
Top