[Closed] New Macs to have Apple's own ARM based processors

However, I am not really comfortable buying a first generation product which may have some rough edges.

If you buy now then when you're next due a replacement they won't be first generation anymore 🙂

Apple have gotten better and better at OS / software updates keeping older machines relevant and useful so you won't be lumbered with something that rapidly obsolesces

I think the decision really lies with the software you use - if you're reliant on certain software then until they've made a decent version for the new Macs then theres no point rushing to that new platform.

First rule of tech - buy what you need now - as there's always something newer/shinier/faster just around the corner! 🙂

10% off at KRCS...
https://www.krcs.co.uk/offers

Didn't apple dump their own processors in favour of Intel in proper machines a number of years back. I seem to recall they tried to make it sound like a really good thing despite it being a really embarrassing climbdown on years of development and marketing.

As above though, buy what you want today, just accept more controlled architecture from apple means much less flexibility for you the user.

Also its not entirely a first gen product, sure its a variation on an existing product but no more than the revised processors in each new gen of iPhone are first gen on release day.

I don't think its for all systems, just some. I can't see them dumping x86 completely just for the lower power machines .

Interested to hear other peoples views on the processor change and how it will effect Mac users.

for 99.99999% of users will make no odds whatsoever.

Didn’t apple dump their own processors in favour of Intel in proper machines a number of years back. I seem to recall they tried to make it sound like a really good thing despite it being a really embarrassing climbdown on years of development and marketing.

Yeah, ARM are clever, clever guys, British too, even if they're Japanese owned now.

But tech is a numbers game and Intel produce about 75%-80% of the CPUs for Personal Computers globally, AMD make most of the rest. I wouldn't be surprised if ARM were behind some of the Intel / AMD tech, but on the face of it they're a new player in the computer CPU game, they'll start behind the major players and even with Apples billions they won't have the budget to catch up.

Maybe Apple are going another way though, maybe they're going down the Chromebook route, not as in a stripped down functionality, but stripped down on-device processing.

Anyway, I wouldn't buy a Gen 1.

Didn’t apple dump their own processors in favour of Intel in proper machines a number of years back. I seem to recall they tried to make it sound like a really good thing despite it being a really embarrassing climbdown on years of development and marketing.

Nope. They dumped the power architecture which was IBM owned. This is the first time they've made their own chips for desktops (been doing it a while for phones and ipads).

The new machines will apparently be launched later this year

The [u]first[/u] new machines will.  I think they said it'll be a 2 year process to fully migrate.

I would deduce from that that something like a new MacBook Air and probably Mac Mini will be released in Arm version this side of Black Friday.  Everything else, posher MacBooks/Pros etc. next year, and the real top end i7/i9 desktops, probably not for another 2 years yet.

I certainly wouldn't rush out to buy just for the sake of switching, unless I had a MacBook almost on its last legs.  Although some will of course be queueing outside with a tent until release day (when they can just order online).

For most it will make 0 difference whatsoever. For 1% they'll probably be able to make use of 12 hour battery life instead of 10 on day 1. And about 0.1% would have to make a decision about some Win x86_64 app that they use, that probably needs some intervention and a new version of W10 to run in a VM to make it work.

This change will be more seamless than the last, and the last was pretty seamless for 99.9%.

i wouldn't worry about a gen 1 chip - apple have a vested interest in making it work, the mac range depends on it - and behind the scenes they will have worked on this for a long time (also it's 2020 - migrating architecture is pretty well understood).

If you ever run virtual machines (for windows/linux) or use Boot Camp then just buy Intel now. Likewise if software you rely on won’t be likely to have new native ARM versions ready to go. Performance of these on ARM (emulating x86) will be pretty horrible.

If you just do browser, Adobe CC, Office then you should be good. ARM will hopefully be nice and quick but with much better battery life.

If you ever run virtual machines (for windows/linux) or use Boot Camp then just buy Intel now.

VM already supported… and Windows announcement expected in the autumn. By the time Arm has made its way through the range (3 years from now?), Arm optimised OS of all varieties will be the norm. But I agree, buy the machine that does exactly what you need now, waiting for the switch to Arm to be complete is a mug’s game. The OP should just buy an Intel based machine and crack on…

Interested to hear other peoples views on the processor change and how it will effect Mac users.

I’ll be buying a fairly hefty spec MBP (8core/2tb/64gb ram/8gb graphics) soon and the announcement hasn’t changed my plans.
I usually sell the old MBP and buy a new one every 3-4 years so as long as the intel one I’m about to buy is performing as it should with the current OS and software I’ll keep to that upgrade cycle.
If and when the MBP goes to Arm I’ll only swap if there are significant gains in the speed of photoshop. That’s the one area that for me time=money.
I remember going from G5 tower to intel MacBook and it being a significant step in speed (productivity) same again going to retina 4 core.

I am getting more into coding using Python and Mac seems > Windows for this

IME it is getting to be more and more the case that I end up having to try that much harder with Windows to get open source dev stuff working. You are often making some vm/docker/wsl fudge. If I needed my own machine for paid work I would go Mac, but if I was a hobby coder I'd go Linux.

If you ever run virtual machines (for windows/linux) or use Boot Camp then just buy Intel now. Likewise if software you rely on won’t be likely to have new native ARM versions ready to go. Performance of these on ARM (emulating x86) will be pretty horrible.

There are ARM versions of Windows, but I gather that legacy software is a problem - they can run in an emulator, but performance is poor. I have an old MBP with Windows running in Bootcamp to run some specialist software that is Windows only. I cannot see that working well on an ARM based Mac. That software is fairly niche so they are never going to bother with ARM versions, so running Windows on a Mac is just not going to work.

I am getting more into coding using Python and Mac seems > Windows for this

Curious about that bit.  When I used Mac, Python was preinstalled but horrendously out of date. and a PITA to make stuff be cross compatible across multiple systems, when some/most would rely on their default shipped installation.

Now I see the same with Java, with 80% of all software using 1 version, but newer versions shipping with all major OS, all of which breaks that "write once run anywhere" code.

I now mostly use Linux, which always gets me the latest version of Python, Go, etc. etc., and has native support for all the extra bits like running things in containers (I assume MacOS must support that natively too?), but then use JupyterLab quite a bit which is crossplatform 🙂  The output of that then works pretty much anywhere.

Ok that's enough off topic... back to Acorn... erm I mean Arm...

dangeourbrain
Subscriber

Didn’t apple dump their own processors in favour of Intel in proper machines a number of years back. I seem to recall they tried to make it sound like a really good thing despite it being a really embarrassing climbdown on years of development and marketing.

As above it was Power chips, and it wasn't really a climbdown. The G3 was good and the G4 and G5 were very good, but the G5 ran too hot for what Apple wanted, and there was no sign of things improving.

Previously they switched from Motorola 68K to Power.

hols2

There are ARM versions of Windows, but I gather that legacy software is a problem – they can run in an emulator, but performance is poor. I have an old MBP with Windows running in Bootcamp to run some specialist software that is Windows only. I cannot see that working well on an ARM based Mac. That software is fairly niche so they are never going to bother with ARM versions, so running Windows on a Mac is just not going to work.

Yes probably not the right tool for that job any longer. If the chips are as good as they say though, performance in QEmu might be pretty decent for basic old apps that have no modern equivalent.

Thanks for the responses everyone. At the moment I would classify myself as a hobbyist but more and more of the network and comms stuff I work with is sprouting API's so much of what I am doing is really trying to make sure I am prepared for changes in my work role.

To be entirely honest I do not need a MacBook - it would just be a nice thing to have and I have saved a few quid during lockdown. Maybe I should just use a Linux VM instead and put the money towards the bathroom re-furb thread.

Funnily enough was reading something a while back that speculated that this might happen. Something about the fact that the newer iPad Pros are REALLY good for video editing due to the ARM chip.

I don’t think its for all systems, just some. I can’t see them dumping x86 completely just for the lower power machines .

The latest Arm cores are actually surprisingly fast. Actually benchmark an iPad Pro and be shocked. Or quoting from the Anandtech review of the new Graviton2 servers available on the Amazon EC2 cloud:

Thankfully, this is not the case for the Graviton2: not only were Amazon and Arm able to deliver on all of their promises, but they've also hit it out of the park in terms of value against the incumbent x86 players.

The Graviton2 is the quintessential reference Neoverse N1 platform as envisioned by Arm, aiming for nothing less than disruption of the datacentre market and making Arm servers a competitive reality. The chip is not only able to compete in terms of raw throughput thanks to its 64 physical cores in a single socket, but it also manages to showcase competitive single-thread performance, keeping in line with AMD and Intel systems in the market.

We're not talking a comparison against a low level Intel chip, these are benchmarks against a £10K Xeon. The high performance Mac Pros are sufficiently expensive already that they could just put 80 cores into one and it would be incredibly fast for that sort of high load but heavily threaded jobs that those users typically run (video processing, rendering, etc).

I am getting more into coding using Python and Mac seems > Windows for this

That's a bloody expensive solution to a simple problem. Sounds like you want a new shiny and are trying to justify it to yourself. (-:

Unless you need a GUI I'd just enable Windows Subsystem for Linux and then install Ubuntu straight into the OS. Job jobbed, running Linux then is as simple as hitting Start and typing 'bash'. And if you're running the 2020 build of Windows 10 you get WSL2 which is a huge improvement over its initial incarnation.

The latest Arm cores are actually surprisingly fast.

It's not so much the raw speed it's that the "power users " I know (and I know loads will use their Mac in a different way to what I am about to suggest but many will) use multiple VMS etc.vms work well if just allowing pass through or emulating a basic chip but complicated chip to complicated chip is always a drag down in performance and lots of software will not be recompiled to arm or it will be a expensive purchase to replace.

but on the face of it they’re a new player in the computer CPU game

They've been around in one form or another since the the mid 80's, they used to be called Acorn Risc Machines

They made the BBC Micro that kick started home computing in the UK in about 1982. They’re about as venerable as a hardware maker gets these day.

I always wanted a BBC micro.

Acorn Risc Machines

I remember poor Steve Furber trying to tutor this software leaning CS student though my design of a 'system on a chip' back in the early 90s. The man has the patience of a saint... and a brain the size of a dwarf planet.

I thought Apple dropped the IBM chipset as IBM was asking for a premium for them but it had to so it could develop new chipset as they are not the size of Intel to be competitive. Though I'm probably wrong.

Be interesting to see how these new macs perform and what price they will be.

So a sooper dooper Raspberry Pi then? 🙂

(Not really)

It’s not so much the raw speed it’s that the “power users ” I know (and I know loads will use their Mac in a different way to what I am about to suggest but many will) use multiple VMS etc.vms work well if just allowing pass through or emulating a basic chip but complicated chip to complicated chip is always a drag down in performance and lots of software will not be recompiled to arm or it will be a expensive purchase to replace.

Nobody doing real work in a VM uses a VM which emulates instructions. The reason they're efficient is that the CPU has direct support for virtualisation so for 99.9% of instructions there is zero overhead.

Arm processors do virtualisation just as well as Intel/AMD.

I always wanted a BBC micro.

I actually GOT a BBC Micro! Ha!

Admittedly it was from a car boot sale in 1994, and I was a teenager, and by that point what I REALLY wanted was a Sega Megadrive. But still, I got one.... 😉

@Kuco IBM wanted money up front from Apple to develop the PPC chips further. Apple didn't see that this was reasonable and went with Intel instead.

Gosh. Macbooks are already expensive. Now it looks like they'll cost an ARM and a leg.

Ha haa ha ha haaaaa

ha

They’re about as venerable as a hardware maker gets these day.

Less venerable than Apple, Intel and especially IBM though who are the other hardware makers mentioned here 🙂

IBM were developing PowerPC to run in massive parallel implementations, for midrange and server applications, while Apple’s plans demanded low power (in both senses, relatively) chips… which Motorola were making, but they also saw themselves as competitors to Apple for the domestic, small business, and largely mobile future… and tried to use chip supply as leverage to change how Apple ran its business. Apple not being reliant on outside chip suppliers has been the work of a generation… and is finally close to completion.

Oh, ‘we’ should have done whatever it took to keep ARM (truly) Cambridge based. Small thinking in government not to find a way of making that happen.

I remember poor Steve Furber trying to tutor this software leaning CS student though my design of a ‘system on a chip’ back in the early 90s. The man has the patience of a saint… and a brain the size of a dwarf planet.

cs.man.ac.uk (keysa) 92-95 - you? Never had a clue what he was on about which says a lot (could’ve been the raving mind)

Edit: raving, mind 😉

I had one of the last PowerBook G4s they made. It was a very different time - Apple were lagging behind in sales and it was clear that PC laptops were rapidly gaining a performance edge. Apple were forced to change.

It was also a PITA to own a PowerPC Mac in the fresh new Intel x86 world. There were plenty of apps that didn't work, or weren't updated for the 'old' chip. Course, we didn't call them "apps" in those days. Also, I remember PPC Macs didn't get all the new features in OS updates. I didn't own an Intel Mac until they'd been out a few years but I don't think there was ever much issue with incompatible PowerPC apps (IIRC you could run PowerPC apps in emulation).

This move seems much more proactive / promising. This is not Apple playing catch up, this is them flexing. I'd guess Apple has a much bigger market share these days, too. The other thing is that I'm presuming the shared architecture will allow iPad apps to run on Mac OS and vice versa. Or perhaps the ultimate goal is to remove the distinction altogether. Either way, plenty of incentive to developers to work on ARM versions of Mac software.

I’d guess Apple has a much bigger market share these days

In PCs/laptops? I don't think they do.

What they have done is taken a huge portion of the high-end sales, which is where all the profit is. I don't think Apple are especially interested in moving high volume/low cost computers, which is how you achieve market share.

Tangential: https://www.reuters.com/article/us-softbank-arm-idUSKBN23T28U

And, again, it’s not just Apple moving (more) to Arm based architecture… expect news from Microsoft in the autumn. Hardware developers will be given a better path than the last two attempts (RT[32bit] & 10[64bit])

you?

owersk

🙋🏻‍♂️

I’d guess Apple has a much bigger market share these days
----------------------
In PCs/laptops? I don’t think they do.

~4% of desktop market share in 2009 and now 18%, apparently. I was surprised it was that much.
https://gs.statcounter.com/os-market-share/desktop/worldwide/#monthly-200901-202005

I just remembered my 12" iBook was a PowerPC G4, 800 MHz – 1.42 GHz

My first Mac, I loved that little thing.

they used to be called Acorn Risc Machines

True, but they've been "Advanced RISC Machines" since getting in bed with Apple decades ago.

~4% of desktop market share in 2009 and now 18%

You need to be very careful about interpreting online usage stats. They don't show unit sales. No doubt, Apple are selling more than they did 20 years ago, but a lot of the internet activity will have switched from Windows PCs to Android phones.

I think a lot of people really underestimate just how pervasive ARM processors are these days. They are in pretty much every piece or electronics you can think of in various forms and made by various manufacturers. Their business model is genius in competing without really competing with big FAB based companies like Intel. ARM doesn't actually make anything, they design generalised chips and then license the designs to companies like Apple, Qualcomm, Samsung, AMD etc. and they customise the designs to suit the application in collabaration with ARM designers. Mostly these enhancements are then rolled into the next generlised design for licensing. This has resulted in them being used for everything from game controllers/mouse/keyboard/mobile phones/tablets/smart TV and speakers/microwaves/printers/fridges/in car systems/satellites.. pretty much everything you can think of short of consumer desktop and laptop PC's and it looks like this is changing, a change that's been a long time coming but pretty obvious for a while.

In a world where complex processors like x86 are becoming impossible to shrink into ever smaller packages (Intel has been stuck on 14nm for almost a decade now) ARM designs are far easier, more efficient in both operation and power usage to take advantage.  As was said above, it's easy to see an Apple or major manufacturer putting together a massively multithreaded ARM processor together (100+ cores per chip) and blowing x86 away completely.

It's also about integrating all the systems and the flexibility that in-house chip design affords. This is not completely new for Apple laptops - for example, my MacBook (Pro) has a dedicated ('T2') chip which handles hardware-level data encryption and a few other bits-and-bobs (audio, face recognition). You can't do that with x86 architecture alone. It makes sense that Apple would extend their control to the main CPU (and presumably consolidate all the separate chips). It'll be smaller, lighter, more efficient. So the battery can then also be smaller, lighter etc.

On the current MacBook Pros, the hardware-level data encryption is thanks to the Intel chips, actually. The T2 is doing finger print recognition and other stuff though.

As an aside... projects I refused to work on at uni were on fingerprint recognition and facial recognition, and specifically algorithms that made use of what Arm were working on. Jump 25+ years and chips based on their work and the software written for them can make Intel stuff look like it's steam powered (DSP & AI especially).

I guess they also get to keep more of the value chain by ditching Intel, whose chips aren't cheap compared to licensing Arm cores.

They've been ramping up their processor expertise for years with the Phones and iPads, so going the full hog on a laptop is probably not that bigger step now.