Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Xcode would work on an Arm chip, and you'd be able to run iOS apps. Android Studio is not Apple's concern... nor Eclipse, but I've never come across an Android developer yet running on Windows or Linux. Anyhow, you'd need Oracle to provide a decent Arm64 based Java JVM.

Java - including V8 - is already on ARM: https://bell-sw.com/java/arm/performance/2019/01/15/the-status-of-java-on-arm/

Eclipse and Android Studio are cross-platform IDEs written in Java.

No, that doesn't mean there are ready-to-go downloads of Android Studio, Eclipse and JDK8 for a hypothetical ARM Mac - but it means that producing them would hardly be the Manhattan Project and if there is sufficient interest they will probably appear.

Makes no sense to run a Arm compiled app in a docker container for development and then recompile to Intel to run on a server.

...when you are developing for iOS in XCode, the app is compiled for x86 and run in an iOS-for-x86 sandbox on the Mac, then when you deploy to an iDevice it is re-compiled for ARM. At best, an ARM Mac will be more suited to that job - at worst, you won't notice the difference.

...for Android, you're mostly compiling to processor-independent Dalvik bytecode. If you are writing native code, then most Android devices are ARM so you'll be testing in QEMU or suchlike under software emulation, and working on an ARM machine is going to be far better than that.

...for Docker, maybe you're one of the subset of developers writing C/C++, in which case developing on Docker for ARM makes as much sense as... well... developing for iOS on an x86 Mac. However, a lot of Docker developers are writing primarily in HTML5, Node.JS, Python, PHP, Java which is all processor-independent, and layering this on top of pre-built containers with the necessary software 'stack' (all the usual suspects like Node, Mongo, MySQL, Apache, NGenix are already up and running on ARM Linux). The alternative, of course, is just spinning up a Docker server on a beige box in the server room, or somewhere in the cloud: there's not much difference between using that and a headless VM and most of the Docker tools are network aware. There's some very interesting remote development stuff in Visual Studio Code now that supports docker and remote Linux systems.

There is a lot of interest in ARM (and other architectures) on servers at the moment and significant effort is already being put into Linux on ARM and easily targetting multi-platforms on Docker (and the Linux/Unix world has always been CPU-agnostic, anyway).

No one is asking you to change your workflow overnight. It is absolutely true that if Apple pushes an ARM transition through too quickly - e.g. by letting the Intel models rot on the vine for two years before launching ARM machines - so people like you are stuffed as soon as you need to replace your current Mac - it will be a disaster. Yeah, being cynical that's pretty much what Apple have a track record of doing with the Mac Pro trashcan and the 2016 MBP. However, lets start with the glass half full and assume that you'll have 3-5 years for your workflow to evolve, and that Apple will recognise the need to work with and encourage key developers to port their products to ARM.

In terms of CPU's, the Apple Bionic System-On-A-Chip's contain CPU and stacked RAM. Do you think they could stack 64GB of RAM in there? No.

Unless Apple are completely stupid, they're not just going to take a chip designed for a passively cooled, ultra-thin, ultra-low-power tablet and stuff it in a high-end "Pro" laptop or desktop.

Here's what an ARM server CPU looks like: https://www.theregister.co.uk/2018/09/18/ampere_shipping/

In summary: it's a 32-core 64-bit Armv8 CPU clocked up to 3.3GHz in turbo mode, with a shared 32MB L3 cache. It supports up to 1TB of DRAM from 16 DIMMS plugged into eight DDR4-2667 memory controllers, has 42 lanes of PCIe 3.0, draws up to 125W, and is a single-socket chip

...except, because ARM is mix-and-match, Apple might instead decide on different features to optimise it as a workstation rather than a server.

What about Thunderbolt 3 support?

Go google USB 4 - which includes TB3 compatibility - the standard was released this summer and actual hardware is expected in a year or two.

That said - I think the ARM can has been kicked down the road a bit with the release of the Xeon Mac Pro (which commits Apple to Intel for another 4 years or so unless they want to look really stupid) - and the way iPadOS has been developing as more of a laptop alternative (personally skeptical about that, but its still happening).
 
Wow, you and I have very different definitions of "there." Windows for Arm is an unmitigated disaster and the vast, vast, vast majority of production systems on the planet are running x86 hardware and containers with Arm's market penetration way down in the low single digits with an indiscernable growth curve.

As an enterprise software developer building software for on prem and cloud deployments, my work product is going to be x86 docker containers (for Kubernetes running on Intel kit) for at least the next five years. One hundred percent of our in house infrastructure is x86. One hundred percent of our cloud resources are x86-based and we're not even talking about experimenting with Arm.

If I can't keep doing that work with macOS then I'm going to have to move to a different platform. macOS on Arm would be the end of the road for me.

That depends on how ARM server take off.
AWS ARM server is already good in price/performance ratio and I'm already using them. In not too far future ARM cloud server will be part of your development. I know for enterprise hardware cost is nothing so they will keep using outdated overprices things for super long time but at the end they will change since big names that owns datacenter will switch and every one will follow.
And I believe Amazon's marketing team is busy working with some key customer to try to start this momentum.

CPU arch should never affect any one that's not using assembly.
 
With Intel already selling its failed modem business to Apple and everyone else transitioning away due to constant delays (Windows 10X, ChromeOS, etcetera) combined with Intel’s probably never getting to 5nm I just don’t see the company surviving much longer...
 
  • Haha
Reactions: ddtmm
a lot of people are very much in favor of arm based macs.
Mostly those that have convinced themselves that Geekbench scores are indisputable, and that there's all of this untapped potential hiding in Apple's CPU designs.
 
Apple still designs ARM chips based on ARM's core technology. Apple didn't invent ARM. And if ARM suddenly decides to stop licensing it's technology, Apple couldn't proceed with anything based on ARM core technology. And yes, AMD (and Intel also) DOES actually use "parts" from each other and slap them together. AMD and Intel has for a very long time (and still has) a full cross-licensing agreement, which enables them to use whatever tech the other company has developed. And they do.
Wrong.

For the umpteenth time: Apple DOES NOT use ARM cores. They make 100% custom designed cores that run the ARMv8 ISA.
 
Mostly those that have convinced themselves that Geekbench scores are indisputable, and that there's all of this untapped potential hiding in Apple's CPU designs.

Based on SPEC estimate, A13 single core @ 5W is faster than 3900X @ 40W let along clock by clock IPC.
SPEC is not a small benchmark that can fit into CPU cache. This is real performance and Intel has been submitting this score officially to their database for more than a decade.
 
  • Like
Reactions: Falhófnir
In terms of CPU's, the Apple Bionic System-On-A-Chip's contain CPU and stacked RAM. Do you think they could stack 64GB of RAM in there? No. They'd have to make the chip die physically larger to be able to add in external RAM etc. What about Thunderbolt 3 support? Making a desktop/server class chip for Mac's tiny single digit marketshare? Not going to happen, the cost is too great for the little return, it just doesn't make any sense.
They could stack 8GB RAM. If Apple does an ARM based Mac, it will be in the 12" MacBook form factor, fanless. I can't imagine how Apple would sell enough to make back their R&D investment, but whatever.
 
Mostly those that have convinced themselves that Geekbench scores are indisputable, and that there's all of this untapped potential hiding in Apple's CPU designs.

...while you have confused "ARM Macs" with "just stick an A12X in a 15" MacBook Pro".

What the benchmarks suggest is that the A10/A12 would make a great system-on-a-chip for a fanless 12" MacBook or MacBook Air replacement, even if they can't sustain those benchmarks over long periods (neither can Intel chips, so its a fair comparison). You don't pick an iPad or MacBook Air for video rendering. The higher-end MBPs, iMacs and Mac Pro would probably keep active cooling to support higher performance CPUs.

The "hidden potential" is a no brainer when it comes to increasing clock rates, sustaining 'turbo boost' for longer, adding more cores or having 8 high-performance cores (the A12 is 4 high-performance cores and 4 low-power cores) where more space, better cooling and bigger batteries are available.

Although the Mac Pro would be the last to go (esp. as it isn't even available yet) there are already ARM server-class chips out there with 32 cores, PCIe and 1TB RAM support that could challenge the Xeon - there are even ARM-based supercomputers that combine ARM cores with vector processing units.

The main point of Apple switching to ARM is that they can design their own pick-n-mix processors to suit the machines they want to build.
 
...while you have confused "ARM Macs" with "just stick an A12X in a 15" MacBook Pro".

What the benchmarks suggest is that the A10/A12 would make a great system-on-a-chip for a fanless 12" MacBook or MacBook Air replacement, even if they can't sustain those benchmarks over long periods (neither can Intel chips, so its a fair comparison). You don't pick an iPad or MacBook Air for video rendering. The higher-end MBPs, iMacs and Mac Pro would probably keep active cooling to support higher performance CPUs.

The "hidden potential" is a no brainer when it comes to increasing clock rates, sustaining 'turbo boost' for longer, adding more cores or having 8 high-performance cores (the A12 is 4 high-performance cores and 4 low-power cores) where more space, better cooling and bigger batteries are available.

Although the Mac Pro would be the last to go (esp. as it isn't even available yet) there are already ARM server-class chips out there with 32 cores, PCIe and 1TB RAM support that could challenge the Xeon - there are even ARM-based supercomputers that combine ARM cores with vector processing units.

The main point of Apple switching to ARM is that they can design their own pick-n-mix processors to suit the machines they want to build.

And stop waiting on Intel's still not yet delivered LPDDR4 support.
MBP2016 was a joke with only 16GB ram. It really reminds me the wait for a G5 powerbook.
 
If Apple does an ARM based Mac, it will be in the 12" MacBook form factor, fanless. I can't imagine how Apple would sell enough to make back their R&D investment, but whatever.

You're talking about an iPad Pro in a different case with a keyboard and trackpad, running MacOS. Not a massive R&D project, c.f. the iPad Pro itself which doesn't exactly sell in iPhone quantities. It wouldn't be a replacement for your 15" MBP, but it would leave a 12" Intel MB in the dust, and the sort of people the 12" MBP (or the current Air) are targeted at probably won't worry about running Creative Cloud or Docker.

I don't think Apple are going that way at the moment - but it would be perfectly feasible. I'd be very surprised if they don't already have MacOS up and running on the iPP in some locked basement in Cupertino.
 
  • Like
Reactions: MikeZTM
Based on SPEC estimate, A13 single core @ 5W is faster than 3900X @ 40W let along clock by clock IPC.
SPEC is not a small benchmark that can fit into CPU cache. This is real performance and Intel has been submitting this score officially to their database for more than a decade.
I'll believe it when I see an Apple CPU transcode a 2hr mkv movie into mp4 faster than a 3900X.
 
I've been hearing that for 15+ years !

It's much more complicated than that !

Hold on, youngsters, while I hike my pants. Why, I remember back in ol' Nineteen Dicketty-Seven, back when Apple released the G3 and the "Think Different" marketing push to compete with Wintel Pentium. Previously, all business writers and many tech journalists had written off Macintosh as obsolete as Coleco.

Mac enthusiasts made a point about the strength of RISC vs. CISC processing as a vector which the PowerPC fan club hoped would close the gap with Intel. I remember the site Mackaido (sp?) was a blog with great engineering hype about this.

What changed was Apple going to OSX, which was a bold move which removed a lot of the need for legacy cruft, especially as the ecosystem of X evolved, and Intel pwning everyone with the Core series.
 
I've been hearing that for 15+ years !

It's much more complicated than that !

It's complicated sure. But Apple has already done a complete architecture change twice and it went really smoothly especially the last time. They're certainly well capable of it. This time it's arguably even easier as their entire development toolchain has been capable of the new platform for years.

I think it'll be a great change if they do it. As long as they will keep offering the full macOS of course. If they go for something more limited (e.g. only app store apps, only Catalyst apps, no terminal access, something like that) it will be useless. I'm thinking something like Surface RT here.

But if they do it well they'll pull it off fine.
 
Last edited:
  • Like
Reactions: Falhófnir
Well, if they came up with Macbook Air style ARM notebook with x2 battery life and 4x time the performance, then they can succeed. Otherwise, it would be a failure and waste of money.
 
That’s just because PCs have big heat sinks and fans. Put a good sized Peltier cooled heat sink on the back of an iPad behind the A12x, and it will run tons longer without throttling.

Peltier coolers have HUGE power requirements.
 
  • Like
Reactions: DanBig
As an enterprise software developer building software for on prem and cloud deployments ...

Apple's laptop market is magnitudes larger than your subset of enterprise developers. They design for that much larger market, where battery life and portability might be more important than not having to recompile docker containers.
 
I'll believe it when I see an Apple CPU transcode a 2hr mkv movie into mp4 faster than a 3900X.

You do not need to transcode mkv into mp4.
Usually mkv contains h264 video stream that can be copy paste into a mp4 package.

If you means encode h264 stream without hardware encoding(aka CPU encoding) then it is already part of SPEC's benchmark suite.


A13 is almost same level of Intel desktop parts in this test.

And modern problem actually already have a modern solution--AISC encoding.
My desktop iMac Pro is already using T2 chip to encode H265 video streams and the quality is absolutely awesome.
 
Last edited:
Well, if they came up with Macbook Air style ARM notebook with x2 battery life and 4x time the performance, then they can succeed. Otherwise, it would be a failure and waste of money.
Nah. A lot of people buy updated PC laptops when Intel ships a chip just 20% to 30% faster. So a 2 pound macOS-on-ARM MacBook just 20% faster than a Core i5 MacBook Pro, and cheaper but with 2x the battery life, would probably sell like hotcakes.
 
  • Like
Reactions: Oberhorst
So do the high end PCs and servers used to produce the top of the x86 benchmark results.

True but for cooling a server they don't strap a Peltier on it and call it good. Many datacenters AC the racks themselves.
[automerge]1573172879[/automerge]
Apple should work with AMD to do a custom chip based on Ryzen like MS did with the Surface Laptop 3. AMD's chips already offer great (if not class leading) performance and have decent (if not great) graphics.

Would be a huge win for everyone. Intel is having chip shortages, high costs, and has for a lot of time produced mediocre processors.
 
You're talking about an iPad Pro in a different case with a keyboard and trackpad, running MacOS. Not a massive R&D project, c.f. the iPad Pro itself which doesn't exactly sell in iPhone quantities. It wouldn't be a replacement for your 15" MBP, but it would leave a 12" Intel MB in the dust, and the sort of people the 12" MBP (or the current Air) are targeted at probably won't worry about running Creative Cloud or Docker.

I don't think Apple are going that way at the moment - but it would be perfectly feasible. I'd be very surprised if they don't already have MacOS up and running on the iPP in some locked basement in Cupertino.
"I don't think Apple are going that way at the moment "

Which way do you think they're going?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.