Mac With Apple-Designed Arm Processor Coming in First Half of 2021

[MagnusVonMagnum]

The magnitude of games not working anymore due to the removal for 32 bit applications on Catalina is not comparable to incompatibilities on Mojave.

If you say so. All the games I've played since upgrading have worked in Mojave just fine. But I keep El Capitan on an external hard drive just in case (plus I can run Windows 10 if need be).

There is a reason the switch was much debated in Mac gaming circles and articles had been written to explain the situation. E.g. you can find a list of 32 bit incompatible games here, although I doubt it's complete.

Apple could have offered some kind of an official VM for older 32-bit stuff without having to purchase an expensive (and frankly incomplete in that sound and graphics acceleration didn't work for me running Snow Leopard Server; I don't know if it's any different for running newer versions of OS X, but without that kind of support, Mac games aren't playable through a VM. Windows support in VMWare has been far better, IMO.

Steam for Mac was released in 2010. Even before that, it's not like standalone games didn't exist on Mac either.

I never bought or had so many games as between 2012 and 2015 or so when OpenGL was still supported and Steam for Mac was running full pace. They even added older games like Knights of the Old Republic that were never available for the Mac before! They even added Dark Forces on Steam for the Mac! (ok that was a bit of a hack, but still). Dragon Age Origins and Dragon Age II were both available. Where's Dragon Age Inquisition for the Mac? METAL and a lack of suitable GPU hardware killed it.

I understand gaming might not be the driving factor for Mac users, but it's still unquestionably a relevant use case.

It's not a driving factor for fanboyz. But they support everything Apple does no matter what. They only ever use Apple apps and so they want everything that Tim wants and never have to worry about incompatibilities. They would never buy or use anything not from the Apple Store. Tim can do no wrong in their eyes. I wish I was kidding....

 

[DeepIn2U]

From the press release -

“Apple has successfully completed the transition to using Intel processors in just seven months—210 days to be exact,” said Steve Jobs, Apple’s CEO. “And what better product to complete it with than the new Mac Pro, the workstation Mac users have been dreaming about.”

Pricing & Availability
The Mac Pro is shipping today with the standard prebuilt configuration, including two 2.66 GHz Dual-Core Intel Xeon processors and priced at $2,499 (US), through the Apple Store® (www.apple.com), Apple’s retail stores and Apple Authorized Resellers. The 20-inch Cinema Display, 23-inch Cinema HD Display and the 30-inch Cinema HD Display are available through the Apple Store (www.apple.com), at Apple’s retail stores and Apple Authorized Resellers for a suggested retail price of $699 (US), $999 (US) and $1,999 (US), respectively.

There’s always shipping times involved for any product bought on the online store. I have zero proof that Apple has the base config in the physical store that day and, frankly, I could care less. Your contention was that the transition from PowerPC to Intel took 18 months, and it did not take that long, at least from a hardware perspective.
[automerge]1582726753[/automerge]
Here’s the bottom line - If Apple has actually made the decision to move to in-house CPUs then it’s done and it’s time to prepare for the transition. I’ve been the first through two and they were remarkably painless.

If you’re married to Intel, then best to buy a new Mac soon so that you can keep the “compatibility” you want or need. Or move to a Windows PC, or cling to what you currently own until the bitter end. Dealer’s choice.

If you’re not married to Intel, then you’re not really all that bothered by this rumor. So, make a bowl of popcorn, pull up a seat and enjoy the spectacle.

Either way, if it’s true, it is inevitable.

Sorry but my contention stands better than you 210 days.

This transition officially (by announcements)
June 6, 2005: Apple announces its plans to switch to Intel processors at the Worldwide Developer Conference and released a Developer Transition System, a PC running an Intel build of Mac OS X 10.4.1 with a Trusted Platform Module[failed verification] in a modified Power Mac G5case

January 10, 2006
15" MacBook Pro and iMac Core Duo line, both using an Intel Core Duo chip

That's your 210 days / 7 months - FYI this did NOT complete the entire product lineup. Mac Pro was not announced in this time frame.

August 7, 2006: "Transition Complete" - Apple announces the Intel-based Mac Pro and Xserve.

14 months. I claimed 18mths originally.

Are you re-writing history or just fitting what suits your rebuttal?

 

[Kpjoslee]

Check Amazon Graviton2 32 cores ARM server processor.
On sepcjvm it's faster than current Xeon core by core!

I don't think that is accurate to say it is faster than Xeon core by core. They compared per-vCPU performance which is only a single thread on Xeon.

 

[Zdigital2015]

Sorry but my contention stands better than you 210 days.

This transition officially (by announcements)
June 6, 2005: Apple announces its plans to switch to Intel processors at the Worldwide Developer Conference and released a Developer Transition System, a PC running an Intel build of Mac OS X 10.4.1 with a Trusted Platform Module[failed verification] in a modified Power Mac G5case

January 10, 2006
15" MacBook Pro and iMac Core Duo line, both using an Intel Core Duo chip

That's your 210 days / 7 months - FYI this did NOT complete the entire product lineup. Mac Pro was not announced in this time frame.

August 7, 2006: "Transition Complete" - Apple announces the Intel-based Mac Pro and Xserve.

14 months. I claimed 18mths originally.

Are you re-writing history or just fitting what suits your rebuttal?

Couldn’t care less...you win, I’m wrong...please feel free to wallow in the past while you argue semantics about **** that doesn’t matter anymore anyways.

Change is inevitable...and good for the soul.

 

[firewood]

But my question has always been - why go through the trouble for a consumer-focused, locked-down ARM-based Mac

My guess is that it won't be. My guess is that this hypothetical product line will be a Pro level Mac, but for Pros who don't need crusty old legacy apps, but who instead value silent operation, long battery life, and more timely hardware revisions. Not locked down, but as open as Pro software developers need. Depending on how well it sells, after a few years down the line Pro software companies will follow the money or go under. Thus helping to kill off legacy software even faster.

 

[dzankizakon]

Still don't think it will be ARM-based, rather, a full-custom Apple design. Stay tuned!

Apple's custom CPU designs are all ARM-based. ARM is a company that created the ARM architecture and its own family of instruction sets. Apple licenses and implements this architecture in their own way, and so do other companies such as Qualcomm, Huawei and many others. Apple would, of course, use their own implementation of ARM architecture in their future laptops - it would be utterly inconceivable for Apple to use other companies' implementation if they switch away from Intel.
[automerge]1582805577[/automerge]
Switching to their own processors would save Apple a lot of money and simplify product development. However, where is the benefit for the end user? I would not mind a MacBook Air with the price of a MacBook Air and the performance of a MacBook Pro, but I doubt Apple would do such a thing.

I am absolutely positive Apple has built prototypes of MacBooks with A processors, but do they know how to position them or sell them? "It's the fastest MacBook Air we've ever built and it's $100 cheaper" - "so what??".

 

[Brammy]

Switching to their own processors would save Apple a lot of money and simplify product development. However, where is the benefit for the end user? I would not mind a MacBook Air with the price of a MacBook Air and the performance of a MacBook Pro, but I doubt Apple would do such a thing.

Battery life is one area I'd expect to notice. Apple's ARM chips would have better power efficiency, and they design the OS and chip for optimum power usage.

 

[MikeZTM]

I don't think that is accurate to say it is faster than Xeon core by core. They compared per-vCPU performance which is only a single thread on Xeon.

Single thread on Xeon is single core on Xeon.
All threads on Xeon are equal. It's just two thread shares same core and resources.

And anyway the price for them are per vCPU so they are cheaper and faster at same time.

 

[Val-kyrie]

There's two ways this can go.... They can cream the current Macs and make everyone drool over the new machines. Or they'll have similar performance as the Arm Windows laptops.

Knowing Apple I think it's been a long time coming and the new CPUs will be really fast, otherwise they'd stick to Intel or expend their x64 lineups with AMD. Apple has made MacOS extremely platform agnostic. We often over estimate the % of people who really care what platform the OS is running. Yes some people would stop buying macs because it won't work with Windows but that's a minority and might not even be an issue since MS has an Arm version if you really care enough about it.

The pinch will come when everyone has to repurchase their software.

What happened to 2020? ☹

Expect the announcement at WWDC 2020

 

[MikeZTM]

The pinch will come when everyone has to repurchase their software.




Expect the announcement at WWDC 2020

For home users, almost nobody need to pay anything. Your daily software are usually just an app in the browser.

And for a lot of graphics pros their software is already subscription based and do not need to pay twice for a update.

Since recompiling current software to new CPU is really easy I do not think anyone will charge a premium just for that.

You can check how fast some open source software adopt Windows on ARM right now, including complex software like Dolphin emulator for Wii.


[automerge]1582829596[/automerge]

Apple's custom CPU designs are all ARM-based. ARM is a company that created the ARM architecture and its own family of instruction sets. Apple licenses and implements this architecture in their own way, and so do other companies such as Qualcomm, Huawei and many others. Apple would, of course, use their own implementation of ARM architecture in their future laptops - it would be utterly inconceivable for Apple to use other companies' implementation if they switch away from Intel.
[automerge]1582805577[/automerge]
Switching to their own processors would save Apple a lot of money and simplify product development. However, where is the benefit for the end user? I would not mind a MacBook Air with the price of a MacBook Air and the performance of a MacBook Pro, but I doubt Apple would do such a thing.

I am absolutely positive Apple has built prototypes of MacBooks with A processors, but do they know how to position them or sell them? "It's the fastest MacBook Air we've ever built and it's $100 cheaper" - "so what??".

Is Apple the company that will only use their own invention to build a new device that is faster than higher tier device and priced lower?

I guess they will build a fan-less mobile workstation that rivals Mac Pro in a lot of use cases.
They already have FPGA afterburner in Mac Pro. Why not having a full ASIC version on a mobile implementation?

Or something that's totally unheard of. Like a screen-less iMac that connect to your glass wirelessly?

And the problem for Apple right now is Intel is dragging them behind. 16GB ram on 2016 MacBook Pro is a joke since Intel can not support more ram on their CPU. This is just like how PowerPC G5 performed back in the days slowing down Apple's laptop lineup.

 

[firewood]

The pinch will come when everyone has to repurchase their software.

Which will be good for many Mac software developers. Significant repurchase revenue opportunities means newer apps and less abandonware. For open source software, or for updates to paid apps or to subscription apps that one has already enabled, there will likely be no repurchases needed.

 

[nickinVictoria]

So dumb. First Catalina. Now this. Nothing worthwhile will run on it.

If it is done like in NeXTStep you won't know the difference. When I was on a HP, Sun, Intel or Motorola the apps all worked except for the speed.... The HP was so fast

 

[cmaier]

If it is done like in NeXTStep you won't know the difference. When I was on a HP, Sun, Intel or Motorola the apps all worked except for the speed.... The HP was so fast

HP’s PA-RISC was a fascinating (and weird) architecture. It’s influence was felt in Itanium, which multiplied the weird by 10.

 

[kyykesko]

If they decide to make it a separate machine in between macbook and ipad it'll fail. If they convert the entire line to ARM it'll mean I'll have to stick with my 2017 MBP until ~2022 as buying the current model doesn't make sense as it doesn't improve things enough for me (even though I do like the new keyboard more, but still not willing to pay a few thousand just for that) and the first gen ARM macbooks will be useless anyway. And the software developers will lag behind for a year or more. I'm talking to you, Adobe & the rest. Plus the SW developers will love having to support two different code bases for the next half a decade or more because the intel macbooks won't all of a sudden disappear anyway.

Anyway, to get people to switch over happily the hardware would have to be something amazing. Give me a 20h+ REAL LIFE battery life and I consider it. That is, in a year or two after the release so all of the software I need works. Rosetta-like virtualization crap killed the battery life the last time so I'm not holding my breath expecting anything out of it. It's native software or nothing at all.

 

[cmaier]

Plus the SW developers will love having to support two different code bases for the next half a decade or more because the intel macbooks won't all of a sudden disappear anyway.

It’s one code base, not two.

 

[nol2001]

Anyway, to get people to switch over happily the hardware would have to be something amazing. Give me a 20h+ REAL LIFE battery life and I consider it.

A 'new' design MacBook with 120fps display would be the kind of thing that would create demand. Some people might want a built in 5G modem too.

 

[cmaier]

A 'new' design MacBook with 120fps display would be the kind of thing that would create demand. Some people might want a built in 5G modem too.

I think most people would prefer doubling the battery life to doubling the refresh rate. Most people can’t even tell the difference in refresh rate.

I figure they’ll come up with an Apple Pencil-compatible device with 5G, and 12 hour battery life.

 

[MagnusVonMagnum]

For home users, almost nobody need to pay anything. Your daily software are usually just an app in the browser.

Wow. You REALLY aren't a computer user. You should just have a phone or iPad.... You clearly have no conception how much software some of us have that are just "home users". NONE WHATSOEVER. I have well over 200 software packages here on my notebook alone.

 

[dspdoc]

Wow. You REALLY aren't a computer user. You should just have a phone or iPad.... You clearly have no conception how much software some of us have that are just "home users". NONE WHATSOEVER. I have well over 200 software packages here on my notebook alone.

It seems theses types are about to potentially create a serious problem for those of us that actually use applications in our work. They’re completely out of touch with the need for applications outside of Safari and Mail. SMH

 

[PostApple2016BrainWash]

A 'new' design MacBook with 120fps display would be the kind of thing that would create demand. Some people might want a built in 5G modem too.

No doubt. A 120hz display is the new standard on every Apple MB LINEUP.

 

[theorist9]

Are there any sophisticated, reputable tests showing how close ARM processors have gotten thus far in competing with X86 desktop processors for typical heavy desktop tasks? For instance: photo and video processing; working with large Excel/Word/PowerPoint documents (using MS Office for ARM); or doing calculations in Mathematica (currently available for ARM at least on the Raspberry Pi, where it's much slower than on desktop chips; not sure if it can be tested on faster ARM chips)?

 

[cmaier]

Are there any sophisticated, reputable tests showing how close ARM processors have gotten thus far in competing with X86 desktop processors for typical heavy desktop tasks? For instance: photo and video processing; working with large Excel/Word/PowerPoint documents (using MS Office for ARM); or doing calculations in Mathematica (currently available for ARM at least on the Raspberry Pi, where it's much slower than on desktop chips; not sure if it can be tested on faster ARM chips)?

A couple of those aren’t really interesting tests. Photo and video processing will depend a lot more on GPU, or on other, non-CPU, cores on the SoC. And Excel/Word/PowerPoint doesn’t really stress any modern CPU.

Mathematica may be an interesting test - does it use FPU, or is it infinite precision?

 

[theorist9]

A couple of those aren’t really interesting tests. Photo and video processing will depend a lot more on GPU, or on other, non-CPU, cores on the SoC. And Excel/Word/PowerPoint doesn’t really stress any modern CPU.

Mathematica may be an interesting test - does it use FPU, or is it infinite precision?

Photo and video processing: Nope. Depends on your task. See, for instance, https://www.pugetsystems.com/labs/a...Force-RTX-Performance-1269/#BenchmarkAnalysis :
"[With Photoshop]...once you get to a mid-range [GPU] card, there really isn't much to gain from using a more powerful GPU....Photoshop is typically much more CPU-limited even when we are using the Core i9 9900K which is the fastest CPU for Photoshop currently available."


Office: Again, no. I get delays and spinning beachballs when performing various operations in those Office applications, if my documents are sufficiently large and complicated (which they often are). And these operations are CPU-limited, rather than I/O limited; when I check my Activity Monitor, the CPU is pegged at ~100% for that application until the operation is completed.

Granted, my CPU is from 2014 (at the time, it was the fastest non-extreme mobile CPU Intel made). But even if current CPUs have single-core speeds 3 x mine (which they don't), that just reduces, say, a 6-second delay to 2 seconds. Still irritating if you encounter it repeatedly.


Mathematica: With Mathematica, you have a choice of specifying whether the calculation is done with machine precision, arbitrary precision, or exact (infinite precision):

1) Machine-precision arithmetic: Fastest, but limited by machine precision for intermediate calculations, which means the precision of the final calculation can be much less than machine precision.

2) Arbitrary-precision arithmetic: You specify the precision you wish Mathematica to maintain for the calculation, and it does this within the software. This requires inputting numbers using arbitrary precision or exact precision

3) Exact (infinite-precision) arithmetic: Mathematica does the calculation while maintaining numbers and physical and mathematical constants as exact symbolic values. This requires all inputs be exact.

 

[cmaier]

Photo and video processing: Nope. Depends on your task. There are, for instance, many Photoshop tasks that doesn't use the GPU. See, for instance, https://www.pugetsystems.com/labs/a...Force-RTX-Performance-1269/#BenchmarkAnalysis.


Office: Again, no. I get delays and spinning beachballs when performing various operations in those Office applications, if my documents are sufficiently large and complicated (which they often are). And these operations are CPU-limited, rather than I/O limited; when I check my Activity Monitor, the CPU is pegged at ~100% for that application until the operation is completed.


Mathematica: With Mathematica, you have a choice of specifying whether the calculation is done with machine precision, arbitrary precision, or exact (infinite precision)

1) Machine-precision arithmetic: Fastest, but limited by machine precision for intermediate calculations, which means the precision of the final calculation can be much less than machine precision.

2) Arbitrary-precision arithmetic: You specify the precision you wish Mathematica to maintain for the calculation, and it does this within the software. This requires inputting numbers using arbitrary precision or exact precision

3) Exact (infinite-precision) arithmetic: Mathematica does the calculation while maintaining numbers and physical and mathematical constants as exact symbolic values. This requires all inputs be exact.

Well, I guarantee you that even the ARM used in the existing iPad Pros will easily handle office applications as well as a MacBook Pro (if given a MacBook pro’s thermal solution, of course. Heat sink and fans).

As for photo and video processing, photoshop does have a lot of legacy cruft that doesn’t exist in more modern applications. And certainly video is unlikely to bound by integer CPU ops in any modern software. But, again, apple’s top mobile ARM, given a heat sink and fan, will do as well as a current midrange MBP.

Mathematica with arbitrary and exact math would be the most interesting cases, but apple’s processors do have single threaded performance comparable to many intel parts.

 

[chucker23n1]

Office: Again, no. I get delays and spinning beachballs when performing various operations in those Office applications, if my documents are sufficiently large and complicated (which they often are). And these operations are CPU-limited, rather than I/O limited; when I check my Activity Monitor, the CPU is pegged at ~100% for that application until the operation is completed.

One doesn’t follow from the other. The CPU could have a thread spinning while polling for an I/O result.