What does this mean for the Mac Pro? Is there even an ARM based CPU that can compete with the high end Xeon chips from Intel?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Hires ARM's Lead CPU Architect Amid Rumors of ARM-Based Macs as Early as 2020
- Thread starter MacRumors
- Start date
- Sort by reaction score
You mean something like this ? https://insidehpc.com/2018/03/new-arm-based-workstation-opens-doors-hpc-developers/
It's unlikely any of the content creation platforms ("Pro" machines) will change any time soon.
What a ridiculously stupid post.
I doubt too many 12 inc MacBook users require boot camp. My guess is we will see an ARM MacBook in 2020 and perhaps MacBook Air. The MacBook Pros will remain Intel based for a bit longer.
If Apple is going to continue to sell me disposable tech, which I most certainly believe they are, the tech itself had better utilize Apple's freakish, awesome proprietary chips.
Intel Macs constantly give me the feeling that the $400 chip upgrade I wish to use is being held hostage inside a $3000 box.
Going full on ARM means that whatever is made available by Apple is all there is and if you don't like it you can use another OS...
Not that that isn't problematic, but at least I can retire from being enraged about this particular angle.
Intel Macs constantly give me the feeling that the $400 chip upgrade I wish to use is being held hostage inside a $3000 box.
Going full on ARM means that whatever is made available by Apple is all there is and if you don't like it you can use another OS...
Not that that isn't problematic, but at least I can retire from being enraged about this particular angle.
Sure. Give ARM-based processors the TDP, cores, clock speeds, and other features of Xeons and they should compete easily. Apple's A series already rival Core processors with passive cooling and low power requirements. If you're not constrained by portability and a battery, ARM-based processors could be workstation-class.
Last edited:
This, or, another option I've not seen suggested much which was not possible for past transitions - Apple designing on die x86 translation into their A-series Mac chips so that existing binaries will "just work"
[doublepost=1561572004][/doublepost]
Absolutely great post, people do not acknowledge this fact often enough.
They don’t have to just have one architecture for their entire Mac product line. They’ll probably offer Apple-powered consumer laptops at some point, since Intel CPUs for fanless laptops are pathetic even compared to the A12X used in the 2018 iPad Pros. The MacBook Air 2018 gets about an 8,000 in Geekbench multi core, the A12X gets 18,000. That’s just the CPU Apple is using in last years iPad Pros - who knows what they’ve got in store for a consumer laptop.
Unless I’m mistaken, MS has a very good version of Windows that is purpose-built for ARM processors. I haven’t used it but I have seen articles suggesting that it works quite well.
Also, as a company that has been reasonably successful (by whatever metric you choose), I would think that they would plan far enough ahead to consider the need for virtualization. I don’t know enough about the different processors to make intelligent statements, however, I would think that there would be a way to translate the more complex instruction sets to a RISC environment. (Efficiency may be an issue, but at least Apple has the benefit of controlling the architecture decisions.)
These leaked benchmark, probably fake and just informed assumptions, were 4 months ago. Imagine what we'll be looking at a year from now. https://bgr.com/2019/03/15/macbook-arm-processors-benchmarks-beat-intel-scores-leak-says/
Are you aware of the reasons for the transition, specifically the performance per watt problems as well that dogged early 2000s PPC chips and roadmaps?
I do agree that Intel's technology at the time was in many ways inferior, but making the switch resolved the above issues and probably helped a great many switchers to dip a toe in the water of Mac OS by resolving long standing issues with Windows compatibility.
Thank you for pointing me in the right direction. Had a quick scan of the Wikipedia page for this and it seems rather good.
Errr how does working on A72 and A76 lead credence to an custom CPU just for Macs ? If he was the lead archiecture on Neoverse N1 and/or E1
https://www.anandtech.com/show/13959/arm-announces-neoverse-n1-platform
then sure. But there is no more credibility for a "Mac" ARM SoC here than yet another iPad Pro focused one. Frankly given iPad has its own named OS now there could be an even bigger 'fork' coming between the "phone" SoC and the iPad Pro SoC than there was before. This doesn't have to loop in the Mac at all. It doesn't completely rule it out but it far, far, from a sure thing.
Economically SoC just for whole Mac line up still doesn't make any sense. Apple could prune off a low end MacBook ( and perhaps the bottom "half" of the laptop line up) , but they are still way, way off doing anything with the upper half of the desktop line up ( even less so now that the Mac Pro is back from Rip van Winkle sleep. ). However, Apple could do that with just retasking iPad Pro SoCs for the job.
There wouldn't be anything particularly Mac there. the MacBook has the same number of ports as the iPad Pro. Mouse drivers , basic keyboard , etc .... all in iPadOS at this point. Basic USB drive ... iPad OS. etc. etc. If get a future SoC more tuned to the expanded gap between iPhone and iPad upper half then there are "spillage' into the Mac space but no grand take-over of the whole Mac line up.
The former stints at Intel and AMD doesn't say anything substantive about replacing x86 anymore than it did when left there and went to ARM in 2009. Going to a substantive ARM ship like Apple probably means going to contribute to what Apple is already doing.
Today?
No. There are 32-core ARM server chips, though, and various groups are building "ARM-based" supercomputers - although tier power usually comes from specialist vector-processing units controlled by the ARM.
4-5 years down the line when your new $6k Mac Pro is obsolete? No reason that there shouldn't be. The ARM could compete with the Xeon by cramming more cores and specialist processing units on a chip.
The more time that passes, the less applications depend on a specific CPU instruction set, the more high-end computing depends on GPUs, vector processors, neural nets etc. rather than x86, the more lower-end Apps will be written with both Macs and iDevices in mind.
Moving the Mac to ARM is perfectly feasible provided its done at a sensible pace and users aren't prematurely forced to switch because Apple's forgotten to update the Intel models...
I think its got a bit less likely with Apple's investment in the Xeon Mac Pro - and based on the other WWDC announcements it looks to me like Apple's plan is to develop iPadOS to gradually replace MacOS from the bottom up.
Uptake is more in the data center for headless servers and not yet so much for desktop workstations.
https://www.gigabyte.com/us/ARM-Server/Marvell-ThunderX2
https://amperecomputing.com/product/
It seems most ARM licensees are interested in ARM for its low power performance, and as such, those that are licensed to design their own ARM chips do not focus on using the ARM instruction set to make any processors that compete directly with Xeon class chips.
However, ARM is just a general design that these firms can use to facilitate their own, so if Apple has decided to do that, everybody better watch out.
Apple owned a part of the PowerPC alliance. Did that drive Apple Mac prices down? Dream on. HAHA!
2019: the year Apple drops support for 32bit applications
2020: the year Apple drops support for all the software in the world
Because chips don’t exist in a vacuum.
How so? Last time around it was Apple choosing between other small vendors, this time Apple is flush with cash and has internal teams for designing microprocessors so that they can control their own destiny. If they're using that strength to match or beat Intel while achieving greater economies of scale by using one kind of chip across all of their devices, that seems a far cry from waiting for Motorola or IBM to crack the case of a 3ghz PowerBook G5.