Hmm maybe another reason to wait on new macs until after 2020.

[DonCarlos]

I am merely a human, a consumer, and not a tech, but if Apple transitions to ARM processors and dumps Intel next year isn’t that like when they dumped PC based Macs to Intel Macs, rendering older macs of little use? Or am I missing something?

So maybe I will keep saving up for a new 2020 mini or ......

Again don’t slam me too hard, I am just thinking this means something, but alas I am just a consumer.

Intel expects Apple to transition Macs to ARM processors in 2020, report says https://www.digitaltrends.com/computing/intel-expecting-transition-mac-to-arm/

 

[Stephen.R]

For some of us that would make the intel models more valuable.

 

[rmdeluca]

I am merely a human, a consumer, and not a tech, but if Apple transitions to ARM processors and dumps Intel next year isn’t that like when they dumped PC based Macs to Intel Macs, rendering older macs of little use? Or am I missing something?

Well, when Apple switched to the PowerPC over 680x0, there was a clear advantage going from the old to the new in terms of speed and future scalability.

When Apple switched from PowerPC to Intel, Intel had finally reached (or was soon to reach) parity with workstation processors (such as PowerPC) with its Core architecture, so there was no serious loss of capability and they got to ride Intel's meteoric architecture and process improvements over the next decade.

While Apple has done some fantastic work on A12X (and recent prior generations), what is not clear is that Apple can match or exceed Intel in the high-end or workstation CPU space. Intel's R&D budget is still larger than Apple's and Intel, while dallying with many pet projects, is still primarily a semiconductor company.

The point being we may see a bifurcation in macOS space - laptops will do fine with ARM but any serious Apple desktops will lag behind their Windows counterparts unless Apple pulls off a miracle (minus Steve Jobs) of scaling their CPUs and GPUs up to 10x their current TDP levels and beating Intel's 10nm and future 7nm tech.

Another problem is we don't know exactly yet where machine learning will take us in terms of coprocessors/accelerators. A12X does some pretty impressive stuff for a phone form factor chip, but where will neural nets be in three or four years? Nobody knows for sure and Apple going proprietary while the overwhelming majority of machine learning work is done in PC space could hurt adoption of any tech they create, no matter how awesome.

Anywho, sorry for the rant. To answer your original question - "maybe." Not "yes" like it was going from 680x0 to PowerPC and not "probably eventually" like it was going from PowerPC to Intel.

 

[morze]

Might be another reason to purchase prior to ARM processors.

There will be a transitionary period which some software won't be supported until software engineers make the time to transition over. That transitionary period may be fast or slow. Sure, for a consumer that uses a lot of the Apple software it'll be fine. But, I work freelance in an industry that preference stability over latest and greatest. I would never set up a job with ARM processors Macs until it can be trusted. Hence, Apple's upcoming Mac Pro will be released with Intel processors. Some jobs I've done recently still use the original trusty Mac Pro towers.

EDIT: Just realised this thread was under Mac Mini. May be still relevant for some.

 

[Ledgem]

I'm starting to strongly dislike that article, because I've seen it posted multiple times as proof of Macs going ARM and it's a crummy article. There are two problems with it:

1) Quoting another website that is talking about private conversations with people who work at Intel. Intel hasn't heard anything from Apple and as far as we know these officials don't have any data to go off of, they're just speculating like you and I. Yet the article headline talking about "Intel expects..." makes it sound a lot more official than that, doesn't it?

2) Marzipan is being quoted as some proof that Macs will go ARM, which doesn't really make sense. We're already seeing the early fruits of Marzipan in the form of Catalyst. Yet what is Catalyst enabling? iOS apps to be ported to x86 Macs... not the other way around. What it's really doing is allowing iOS developers - who I'd guess outnumber Mac developers - easier access to the Mac, which should theoretically also increase the number of programs available on the Mac, which could in turn revitalize the Mac ecosystem. It doesn't mean that Macs aren't going ARM, but it's definitely not proof that Macs would need to go ARM, given that Catalyst is already happening with x86.

Apple will need to tread incredibly carefully if they want to switch away from x86. Aside from bringing development in-house, it's not entirely clear what benefits they would gain from ARM, either. Their MacBooks are already class-leading for battery life and size, and performance is surprisingly good - I used one as my main computer for a year, even running Capture One Pro and Final Cut Pro on it. Desktops would probably have zero benefit, and all at the risk of alienating other developers and mainstream software that many Mac users currently rely on. And while I can't speak for everyone, much as I enjoy my iPhone and iPad, my computer is still my #1 device. If I'm no longer on a Mac, the strength of having multiple Apple devices that speak to each other greatly diminishes. Microsoft's Surface is a compelling iPad replacement, particularly if my main computer becomes a Windows computer, and while there's no longer a Windows phone, Android is... well, I'd probably stick with the iPhone

But the point is, Apple would need to make a very compelling case for the architecture switch, one that would give enough benefit to their consumers that the negatives would be outweighed. I don't have any special behind-the-scenes information to back up my opinion, but based on keeping up with technology news, I just don't see the benefit nor the appeal beyond making things easier for Apple's management. That's a poor proposition.

 

[ivnj]

But ARM is slower isn't it?

 

[velocityg4]

Well, when Apple switched to the PowerPC over 680x0, there was a clear advantage going from the old to the new in terms of speed and future scalability.

When Apple switched from PowerPC to Intel, Intel had finally reached (or was soon to reach) parity with workstation processors (such as PowerPC) with its Core architecture, so there was no serious loss of capability and they got to ride Intel's meteoric architecture and process improvements over the next decade.

While Apple has done some fantastic work on A12X (and recent prior generations), what is not clear is that Apple can match or exceed Intel in the high-end or workstation CPU space. Intel's R&D budget is still larger than Apple's and Intel, while dallying with many pet projects, is still primarily a semiconductor company.

The point being we may see a bifurcation in macOS space - laptops will do fine with ARM but any serious Apple desktops will lag behind their Windows counterparts unless Apple pulls off a miracle (minus Steve Jobs) of scaling their CPUs and GPUs up to 10x their current TDP levels and beating Intel's 10nm and future 7nm tech.

Another problem is we don't know exactly yet where machine learning will take us in terms of coprocessors/accelerators. A12X does some pretty impressive stuff for a phone form factor chip, but where will neural nets be in three or four years? Nobody knows for sure and Apple going proprietary while the overwhelming majority of machine learning work is done in PC space could hurt adoption of any tech they create, no matter how awesome.

Anywho, sorry for the rant. To answer your original question - "maybe." Not "yes" like it was going from 680x0 to PowerPC and not "probably eventually" like it was going from PowerPC to Intel.

While there was a clear technical advantage for PowerPC. Software at the time was generally quite expensive. As there weren't online updates it required all new software. Otherwise running 680x0 software was like downgrading to a 68020 Mac.

Also most software wasn't PowerPC ready when Power Macs first came out. For those people PowerPC was a downgrade for a while.

Çonnectx SpeedDoubler was a huge help with 680x0 emulation. I remember it made my 7100/66 AV run 680x0 code at 68040 speeds.

 

[Partron22]

Macum non nocere.

 

[filterdecay]

i could see the t2 chips being replaced with some arm chip working hand in hand but so much legacy software would need time to transition or arm would need to brute force x86 code.

 

[throAU]

For some of us that would make the intel models more valuable.

By 2020, i would suggest that ARM will likely be able to emulate intel x86 at the speed of a dual core 2019 machine if the rate of improvement over the past 3 years is any indicator.

Intel are stagnant. ARM are making huge strides in processing power per watt.

 

[filterdecay]

By 2020, i would suggest that ARM will likely be able to emulate intel x86 at the speed of a dual core 2019 machine if the rate of improvement over the past 3 years is any indicator.

Intel are stagnant. ARM are making huge strides in processing power per watt.

I think in the next 10-15 years everything will need to transition off of x86. 10nm now and then 7nm. How much farther can we go until we are talking antman.

 

[Stephen.R]

t2 chips being replaced with some arm chip

The T2 is an ARM chip, it's a variant of the A10.

This has been my theory for a while: an Arm co-processor/SoC that takes on more and more responsibilities.

By 2020, i would suggest that ARM will likely be able to emulate intel x86 at the speed of a dual core 2019 machine

Great, so we'll just need a 14-CPU Arm system to run software from a single-CPU 2019 Mac Pro?