Apple Expected to Announce Arm-Based Mac Plans at WWDC, Transition Away From Intel to Begin in 2021

[MikeZTM]

I am and I don't know any cmaier on Macrumors. Sorry.



Fascinating, you can google the papers I, Jim Keller, have written.



Indeed, it is well worth our time as big-time CPU designers to hang out on Macrumors forums, all day, for a decade.



Someone mislead or lied on the internet? What. That's unpossible.

His posting are well educated and make sense from me as a software developer.

That point alone is enough regardless of he is the designer of K8 or not. Though that's a plus and he did told us some really interesting behind the scene stories -- that make sense by them own.

If Jim Keller himself saying bad about ARM (as a instruction set and end product) I will also be laughing at him no doubt. And I can bet he will not say something like that.

 

[cmaier]

His posting are well educated and make sense from me as a software developer.

That point alone is enough regardless of he is the designer of K8 or not. Though that's a plus and he did told us some really interesting behind the scene stories -- that make sense by them own.

If Jim Keller himself saying bad about ARM (as a instruction set and end product) I will also be laughing at him no doubt. And I can bet he will not say something like that.

Plus I was one of the designers of K8, working for Fred with Cheryl, Bill, Steve, Marius, Tom, Regina, Larry, Michael, Ramsey, Allan, Morgan, and, by the time we were done, a couple dozen other people. (Jim Keller included - he worked on the architecture for the hypertransport bus).

 

[MikeZTM]

Plus I was one of the designers of K8, working for Fred with Cheryl, Bill, Steve, Marius, Tom, Regina, Larry, Michael, Ramsey, Allan, Morgan, and, by the time we were done, a couple dozen other people. (Jim Keller included - he worked on the architecture for the hypertransport bus).

Well I guess someone have to be work with real person himself to fake his identity to this level -- at by that he could just use his name instead since he should be a core engineer anyway.

Thank you from a Athlon 64 3000+ ClawHammer 754 user.

 

[cmaier]

Well I guess someone have to be work with real person himself to fake his identity to this level -- at by that he could just use his name instead since he should be a core engineer anyway.

Thank you from a Athlon 64 3000+ ClawHammer 754 user.

I actually built that one! (I think. We never knew the marketing names, but clawhammer and sledgehammer were ones I worked on, and I actually performed the last minute hack, in vi, to edit the layout to fix a bug for tapeout on that. Which reminds me that I forgot Dhiraj should definitely have been on the list, too. And James. More names will come to me)

 

[Jury]

ARM-Based Macbook will herald a better price that poised to sell twofold in volume.

 

[Richdmoore]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

I could see a longer transition than the PPC to Intel era, due to apple updating the Mac Pro last year, but honestly support for non-consumer focused models probably isn’t the best benchmark.

I guess Apple could also just do a “clean break” as well, like they did with the Apple II to Macintosh transition (and continue to sell both for awhile.). That would be way riskier, and more confusing to the consumers.

 

[cmaier]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

I could see a longer transition than the PPC to Intel era, due to apple updating the Mac Pro last year, but honestly support for non-consumer focused models probably isn’t the best benchmark.

I guess Apple could also just do a “clean break” as well, like they did with the Apple II to Macintosh transition (and continue to sell both for awhile.). That would be way riskier, and more confusing to the consumers.

1. 18-24 months
2. 24-36 months

 

[arkmannj]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

I could see a longer transition than the PPC to Intel era, due to apple updating the Mac Pro last year, but honestly support for non-consumer focused models probably isn’t the best benchmark.

I guess Apple could also just do a “clean break” as well, like they did with the Apple II to Macintosh transition (and continue to sell both for awhile.). That would be way riskier, and more confusing to the consumers.

I think this would depend on among other things if Apple has some sort of “Rosetta” equivalent and xcode ready to natively compile MacOS apps to ARMA or not. I doubt the later will be in question much since they obviously already had a major head start on that with iOS.

having something like Rosetta can greatly accelerate their timeline since they don’t need both users and developers to cut right over. If it can work well without a huge performance penalty then the transition can be a huge benefit in their corner.

I think Apple releasing a new MBP with updated graphics today may suggest that they are ready to support Intel machines for a few years. I think we will see a couple versions of MacOS where intel machines will be barred from some “new features” then cut off. If I had a guess I’d say 2 years with mostly full support (but if you move to ARM you can get new feature xyz) then 1-2 years of reduced support. and then 1-2 years-ish of security type fixes.

of course all of this also does not take into account when Apple stops selling new Intel based hardware. If for some reread on they are still selling new intel machines in 2 years from now that would extend the timeframe... at least for those product models.

 

[MikeZTM]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

I could see a longer transition than the PPC to Intel era, due to apple updating the Mac Pro last year, but honestly support for non-consumer focused models probably isn’t the best benchmark.

I guess Apple could also just do a “clean break” as well, like they did with the Apple II to Macintosh transition (and continue to sell both for awhile.). That would be way riskier, and more confusing to the consumers.

Just reminds you Apple even refreshed the top-of-the-line PowerMac G5 after the Intel announcement.

 

[Alan Wynn]

1. 18-24 months
2. 24-36 months

While I hate to disagree with someone who has as much invested in his fake internet persona (BTW, how much did you have to pay to get IEEE to add your name to those JSSC articles?),I think that the last transition is a good benchmark.

1. 18 months.
2. 36 months.

Less confident in the second number. Unlike the Intel transition, this time they have a great many people who are already experienced with ARM so things should be easier. My biggest question for you as a chip designer, is do you expect the MacBooks to be SoC based or will they do a version of the A-series with a separate North bridge/South bridge?

As a follow up to that question, do you expect to still see discrete graphics chips in the laptops or just Apple's own GPU?

 

[cmaier]

While I hate to disagree with someone who has as much invested in his fake internet persona (BTW, how much did you have to pay to get IEEE to add your name to those JSSC articles?),I think that the last transition is a good benchmark.

1. 18 months.
2. 36 months.

Less confident in the second number. Unlike the Intel transition, this time they have a great many people who are already experienced with ARM so things should be easier. My biggest question for you as a chip designer, is do you expect the MacBooks to be SoC based or will they do a version of the A-series with a separate North bridge/South bridge?

As a follow up to that question, do you expect to still see discrete graphics chips in the laptops or just Apple's own GPU?

Seems like you are more or less agreeing, not disagreeing, with my fake person.

There is a lot of uncertainty.

As for whether they do separate bridges, dunno. For laptops or any other sealed boxes (mini), doesn't seem really necessary, especially since the rumor is they are doing their own GPU, too. I believe that apple thinks that in an ideal world, all the silicon would be in a single chip package.

Also I'm not in a position to guess what they do for ports.

 

[Alan Wynn]

Seems like you are more or less agreeing, not disagreeing, with my fake person.

That mostly refers to the first number. I would be surprised if it took longer than through the end of 2021 for them to stop selling Intel hardware (which means, depending on how one defines "transition begins") could make it as short as 12 months.

There is a lot of uncertainty.

MacRumors quote of the day!

For laptops or any other sealed boxes (mini), doesn't seem really necessary, especially since the rumor is they are doing their own GPU, too.

That would be my expectation as well. Their model of separately powered components on (in?) the SoC (like the M-series) chips would make things cheaper and easier as they could ditch the T-series as a separate part.

I believe that apple thinks that in an ideal world, all the silicon would be in a single chip package.

What I wonder is more on the Mac Pro side of the world - will it be a single multi-core A-series, or a set of A-series SoCs connected together? Will they build a different NUMA model, with 8GB or 16GB of RAM on chip and the rest separate (that would make 3 memory zones: local, shared external, on the other SoC).

So many questions.

Also I'm not in a position to guess what they do for ports.

My understanding is that they are going to move to a proprietary ADB-3, A-SCSI (Apple SCSI), and A-Centronics connectors on the laptops. The desktops will move to the ADC-2 with support for three-phase power.

 

[Sydde]

You forgot the proprietary wiA and ATooth wireless protocols, supporting AppleTalk-X networking.

 

[Alan Wynn]

You forgot the proprietary wiA and ATooth wireless protocols, supporting AppleTalk-X networking.

Sorry, I considered them so obvious that I did not think I needed to list them! You are right however, in the cause of completeness they should be included.

 

[anshuvorty]

In 1998, Apple released the first iMac, powered by a 750. In '02, the "moose hoof" iMac ran on a 7450. In '04, the iMac put all its hardware behind the screen, balanced on a slender, elegant foot and powered by a 970fx. The first Intel-powered iMac appeared in '06, looking very much like its immediate PPC predecessor but with "G5" dropped from the name.

And the PowerBook name goes all the way back to the early '90s – I had one that ran on a 68030.

Thanks for pointing out my ignorance. As you can tell by my post, my knowledge of the Macintosh extends from the early 2000s to the present day.

My apologies for not knowing these facts.

 

[smulji]

There's no growth in the Macintosh, it's forever at 10%-ish marketshare

Market share is just a reflection of new sales for the current quarter or year. What matters even more is user base and the Mac user base has been growing the last 5 years. In a fairly recent financial conference call Tim Cook mentioned that the Mac user base is 100 million users. That's pretty good.

 

[Sydde]

There seems to be some concern that that large user base is carried on the back of the Intel architecture. We shall see. As I recall, the 68K->PPC transition took rather a long time, beginning almost 30 years ago, with the last 68K machines appearing in '94 or '95 and the last 68K-compatible OS about 3 years later (8.1, I think). Then there was a second barrier, beyond which the 4-digit-numbered PowerMac models were not able to run OS X. Intel Macs started on Tiger (10.4) and about 3 or 4 years later, Snow Leopard (10.6) was Intel-only. So, I would guess that if you buy an Intel Mac today, it will probably still be entirely usable in '25, but the ARM machines will be availed of some special features* that you cannot get on the Intels


*(yes, "features", which is a classic synonym for "bugs")

 

[diddl14]

Wouldn't be surprised if Apple announces the first ARM based iMac at WWDC with more or less immediate availability.

They basically just need to rebuild all App Store apps using BitCode to have tons of apps available on day 1. And with hardware supported x86 emulation, basically all existing apps could still work. Maybe even faster than on Intel if the Apple silicon team continued their magic on multi-core design...

 

[dooyou]

I hated the first Intel Macbooks (Pro), so maybe its the same with the first ARM Macbook.

 

[aednichols]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

1. 12-18 months from this Monday
2. 48-60 months. Modern Apple supports systems for quite a long time, and the clock runs from when they stop selling them. That’s another good reason to transition the whole line ASAP.

 

[theorist9]

Once the transition begins, anyone want to predict how long it will take to:
1. Transition the entire Mac line to A processors?
2. Stop issuing new MacOS version updates (other than security/minor updates) for Intel machines?

I could see a longer transition than the PPC to Intel era, due to apple updating the Mac Pro last year, but honestly support for non-consumer focused models probably isn’t the best benchmark.

I guess Apple could also just do a “clean break” as well, like they did with the Apple II to Macintosh transition (and continue to sell both for awhile.). That would be way riskier, and more confusing to the consumers.

1) No idea.
2) It varies by machine and year but, roughly on average, Macs have been able to run the current OS for up to 7 years after purchase. Apple then continues security updates for each OS for two additional years, meaning you can run a supported OS for about 9 years after purchase. I would be very surprised if Apple cut this to just a few years on its last Intel machines. One option, if Apple wanted to transition rapidly to ARM-only OS's, without leaving relatively new Intel machines unsupported, would be to extend security support on the deprecated OS's. So instead of 7 years + 2 years, they could do, say, 3 years + 4 years (which would at least give 7 years of total support).

 

[Sydde]

They could also implement some sort of logic board exchange program that would allow Intel Mac owners to switch their hardware to ARM. They have not done that in the past, but this situation seems somewhat different from previous architecture transitions.

 

[cmaier]

They could also implement some sort of logic board exchange program that would allow Intel Mac owners to switch their hardware to ARM. They have not done that in the past, but this situation seems somewhat different from previous architecture transitions.

How is this situation different?

 

[Tmak74]

I hope they only do this for low-end macs. I use too many virtual machines for this ever to work for me professionally.

I don't think the arm chips can match the performance of the intel H series of mobile processors like the ones in the 16 inch. I have no qualms about buying now and am planning to do so. I'm not going to be the guinea pig on a new processor. Plus the 16-inch base model is 300.00 off and am going to take that deal. Even better I'm able to get it for that price with apple care plus included so I'm set for three years! And if arm macs are gimped somehow with running certain software then Windows 10 here I come. I love both operating systems and at this level of maturity, its mac that's actually stagnating so choices are out there if you have an open mind. I am more inclined to think the high-end laptops will roll with arm much later so I think you're more than good.
[automerge]1592370329[/automerge]

I always assumed the fanless MacBook would be the perfect candidate for an ARM Mac but perhaps Apple really wants to show this is for professionals so will announce something much more powerful, I wonder if they will keep the same names or if the ARM Macs will get new names.

Arm still has a way to go to match intels laptop mobile pro chips. People are really giving apple to much credit here as if they are doing it because of efficiency etc. They are doing it because they will control everything and make more money that's it. I go back and forth between windows and mac os so for me its no big deal but suspect the mac line will become more limited in the future with the Ipad becoming more pro. Apple seems more invested in IOS then Mac OSX. Arm macs will make good ultrabooks. Technology is not mature enough yet to have that kind of raw power without a fan. Also, where is FCPX for the ipad pro if it's so fast? All the youtubers out there are touting lumafusion ?? Apple had ample opportunity to get FCPX out there for the Ipad Pro but they did not. It tells me it's not powerful enough to handle that kind of pro software at this stage. Next year the Ipad Pro could be different with more horsepower and more ram.

 

[cmaier]

I don't think the arm chips can match the performance of the intel H series of mobile processors like the ones in the 16 inch. I have no qualms about buying now and am planning to do so. I'm not going to be the guinea pig on a new processor. Plus the 16-inch base model is 300.00 off and am going to take that deal. Even better I'm able to get it for that price with apple care plus included so I'm set for three years! And if arm macs are gimped somehow with running certain software then Windows 10 here I come. I love both operating systems and at this level of maturity, its mac that's actually stagnating so choices are out there if you have an open mind. I am more inclined to think the high-end laptops will roll with arm much later so I think you're more than good.
[automerge]1592370329[/automerge]

Arm still has a way to go to match intels laptop mobile pro chips. People are really giving apple to much credit here as if they are doing it because of efficiency etc. They are doing it because they will control everything and make more money that's it. I go back and forth between windows and mac os so for me its no big deal but suspect the mac line will become more limited in the future with the Ipad becoming more pro. Apple seems more invested in IOS then Mac OSX. Arm macs will make good ultrabooks. Technology is not mature enough yet to have that kind of raw power without a fan. Also, where is FCPX for the ipad pro if it's so fast? All the youtubers out there are touting lumafusion ?? Apple had ample opportunity to get FCPX out there for the Ipad Pro but they did not. It tells me it's not powerful enough to handle that kind of pro software at this stage. Next year the Ipad Pro could be different with more horsepower and more ram.

The iPad Pro does not have a lot of RAM, doesn’t come with a keyboard, and not a lot of people are looking to run FCPX on it yet. Soon.

But in any event, your theories about processors are wrong. When the MBP with Arm comes out, it will beat the comparable Intel MBP by 20% for most workloads.