Apple Silicon: The Complete Guide

[DanBig]

Not sure the RISC CISC argument is correct.

Intel x86 Processors – CISC or RISC? Or both??

The argument between CISC architecture and RISC architecture is longstanding. For compiler designers, RISC is a little burden since the same...

sunnyeves.blogspot.com

Correct! CISC now has some of the benefits of RISC. Neither offer the needed performance levels for the next gen CPU architecture. But that goes beyond the issue here.

ARM natively supports both canonical and non-canonical code so jumping to it is not as hard as it was for Apple to jump from PowerPC to Intel. The only reason it wasn't that noticeable was the huge performance difference between the two chips!

Apple's ARM design will end up being more a CPU than an APU. Shedding the onboard RAM as well as the GPU logic to standalone chip or offering a dual GPU design. The ultra thin carrier the A-Series design uses will also be shed for a more effective heat transfer design.

By the time Apple produces the chip the performance difference won't be as large as what we have now with Intel if you compare with a 10 nm chip running macOS. It's only going to be faster with a lighter weight OS like iPadOS.

 

[Unregistered 4U]

I suppose I'm coming at it from a video editors perspective, so I'm an outlier as far as use goes,

You’re right, your use is an outlier. For the vast majority of users, them running the Geekbench test is likely the MOST stress they ever have AND ever will ever put on their system!

BTW, the A13 has made major strides towards improving peak performance, and that’s in a form factor as constrained as an iPhone. Add a more friendly thermal environment, and the peak performance can be maintained just as well as with any thermal controlled processor.

 

[DanBig]

So much wrong here it’s not even funny

Apple doesn’t rely on ARM designs. That would be Qualcomm, Huawei and now Samsung (who laid off a bunch of engineers and are no longer making custom cores). They’re the ones who “patch together ARM parts” to make processors.

Apple designs 100% custom cores that are compatible with the ARMv8 instruction set. The same way that AMD designs 100% custom cores that can run x86. Do you think AMD simply uses Intel “parts” and slaps them together? Ridiculous.

I don’t know why people keep perpetuating this myth that Apple just “slaps together” ARM designs to make a processor.

They don't! We always talk about it as ARM architecture not the companies product Arm. Likewise Samsung Exynose chips are based on the ARM architecture. And as you stated the Qualcomm chip is an Arm chip.

Yes, its confusing! As the case of the word defines the context. ARM vs Arm.

 

[MoXoM]

MaciOS

 

[Unregistered 4U]

And the same wouldn't apply to the Turbo Boost of Intel chips?

I don’t think it’s the same as Turbo Boost because the speed of Apple’s processor doesn’t change. All cores running flat out are as fast as it gets. And, once you get to that point, the speed is sustained until thermal limits are hit. With the A13, those limits have moved out a bit, but ARE still there due to being in a thermally constrained space.

 

[johannnn]

What is often forgotten when measuring performance is the length of time the process takes. Short benchmark test is like a drag race between two muscle cars on a flat straightaway. But most Pro's are processing long process activities (video rendering as an example). This is more like a race up a mountain! This is not tested with benchmarks effectively!

Long processing sessions can kill the APU/CPU or GPU device. Have we forgotten all of the failed GPU chips in the older MacBook Pro's?? As we pushed them harder than what they where designed to support. I don't blame AMD or NVIDEA here as they where caught by more aggressive game apps (excluding NVIDEAs bad chip)

So is an ARM based MacBook Possible? Yes! But is it going to work at the needed level of performance? Sadly No, for the more process intensive apps Pro's use.

Then there's the OS elements. iPadOS has started to migrate into a more effective OS for tablets (still has a bit more to go). Can it be used as the basis of a clamshell system Yes! And thats what Apple needs to embrace!

What I foresee is a iPad clamshell filling in the MacBook slot. Thats the market Apple needs to enter. They are almost there now with the iPad Pro line with the keyboard covers. Many people just don't like the cover keyboards and the lack of a trackpad also forces people to touch the screen. A MacBook type of keyboard with a touchpad LCD surface with a full display above would be the ticket here!

What I want for a Pro level system is performance! To render hours of video without melting down. This is where Intel or AMD Ryzen or even the coming ThreadRipper chips make more sense.

But lets look at this a bit differently. What is holding us back in processor design? We are still dependent on single threaded processes. While many apps are multi-threaded they are limited by the scaler architecture of the CPU and then if you push for massively multi-threaded (SMT) across multiple CPU's the collective cores need to run at slower clocks because the chip can't handle running too many cores concurrently on the same die as the thermals would over stress the silicon! So dense dies need to scatterplot the running cores to balance the thermal load across the die this also slows the running process. Neither CISC or RISC architectures are able to get us to this next level of performance!

Is there any data that shows that Apple’s A12/A13 chips are bad at “race up a mountain”? I’m a newbie but I understand that a drag race and a race up a mountain is different. I’m just curious to see some data.

 

[kemal]

(such low hanging fruit)
If Apple release an arm based Mac, they would probably want, like, $1200 for just the arm. And it wouldn't even be VESA compatible.

 

[smulji]

The entire article doesn't mention 'Swift' at all which is really surprising. It is not about 'will they make the transition' but when. They are putting all the little bits and pieces into place to make this transition happening and have been planning it for years. They already have all software vendors programme their software in Swift allowing them to release their software on all platforms at the same time. When the time is right, they just turn on the switch for this transition.

Most big software vendors like MS, Adobe, Autodesk, etc., are not developing their Mac software in Swift. They're using AppKit / Objective-C.

 

[Nugget]

You'll just bootcamp/wine/vmware these apps with Windows on Arm on the Mac. Are these Apps available in 32 bits?

For starters, macOS Arm means the end of VMware/Parallels/Docker/Other hypervisors. All of those rely entirely on the virtualization functions baked into in the CPU. No Arm CPU is going to be able to offer x86 virtualization support that's necessary to run virtual machines on the macOS system. Arm on macOS would the end of the road for virtualization.


Secondly, where would I purchase a copy of Windows on Arm for your theoretical Boot Camp solution? That's not a product that exists for sale in the retail market. There's zero reason to expect that "Windows for Arm" will ever be a shrink-wrapped product that a consumer can just buy and install on their own hardware -- Mac or otherwise.

Lastly, even if I could buy a copy of Windows for Arm, and the app I really care about is 32 bit, the performance would be horrible.

I can't imagine anyone choosing to endure that kind of mess when they could just jump ship to a Linux or Windows based workflow on commodity hardware. No matter how much you may personally dislike Windows, at some point you just have to suck it up to get your work done if macOS goes too far off into the weeds.

 

[Unregistered 4U]

Is there any data that shows that Apple’s A12/A13 chips are bad at “race up a mountain”? I’m a newbie but I understand that a drag race and a race up a mountain is different. I’m just curious to see some data.

With its current use cases, we’ve ONLY seen A series performance in iPhones and iPads where it’s hard to keep them cool. In those cases, the CPU will slow itself after awhile to keep your phone or iPad from getting too hot. What we haven’t seen, is an A series CPU, cooled properly, under sustained load. Some testers did see sustained high performance when running benchmarks in a freezer or sitting on an ice pack, but that’s not really considered “cooled properly“

 

[BigBoy2018]

For starters, macOS Arm means the end of VMware/Parallels/Docker/Other hypervisors. All of those rely entirely on the virtualization functions baked into in the CPU. No Arm CPU is going to be able to offer x86 virtualization support that's necessary to run virtual machines on the macOS system. Arm on macOS would the end of the road for virtualization.


Secondly, where would I purchase a copy of Windows on Arm for your theoretical Boot Camp solution? That's not a product that exists for sale in the retail market. There's zero reason to expect that "Windows for Arm" will ever be a shrink-wrapped product that a consumer can just buy and install on their own hardware -- Mac or otherwise.

Lastly, even if I could buy a copy of Windows for Arm, and the app I really care about is 32 bit, the performance would be horrible.

I can't imagine anyone choosing to endure that kind of mess when they could just jump ship to a Linux or Windows based workflow on commodity hardware. No matter how much you may personally dislike Windows, at some point you just have to suck it up to get your work done if macOS goes too far off into the weeds.

Agreed! And then consider this scenario: Apple moves to ARM, then a year or so later intel leapfrogs ARM in performance and efficiency.
All the ARM cheerleaders in these threads be like: “well dang”

 

[BOOMBA]

This was helped massively by Intel chips being a lot faster than PowerPC.

Are you sure? I distinctly remember Intel chips looking like literal snails compared to PowerPCs.

Another advantage of ARM and a custom chip is that Intel is $crewing everyone on price. The switch could easily shave $500 off the price of a Mac laptop making them much more price competitive while still keeping margins

Umm... If there's one thing Apple has a history of, and Tim's Apple in particular, it's screwing everyone on price. There is no way An ARM MacBook would be cheaper than a current one, and odds are they'll charge more for it because its a new thing.

 

[Cyprusian]

When/if Apple decides to move away from Intel processors it's reasonable to expect it to deploy a robust change management strategy (including effective communication) to support its developer and user bases, while safeguarding sales of Mac computers before, during and after the transition.

In the absence of official Apple announcements, I guess everything lives in the rumour zone which is naturally the raison d'être of this site.....

 

[Rigby]

I think the Surface Pro X exists because Microsoft wants to cultivate this potential future without Intel, and Lakefield exists because Intel knows Microsoft’s future roadmap and they will either have a product that fits into it OR Qualcomm will. If Intel can hit their targets, Microsoft could keep ARM and Intel going side by side indefinitely.

Microsoft will only keep the ARM version of Windows going if ARM-based Windows devices gain larger traction with hardware OEMs. They will not do it just for their own device.

If, as has happened in the past, Intel misses their release windows year after year, then Microsoft has an ARM loaded future, just ready to pull the trigger on it.

That was mostly because they had issues with their 10nm process, but that seems to be solved. And I'm not sure if depending on Qualcomm is any better than depending on Intel ...

 

[Nugget]

Are you sure? I distinctly remember Intel chips looking like literal snails compared to PowerPCs.

PowerBook G5 confirmed!

 

[BOOMBA]

Interesting, especially as I really don’t see how you got to that view, when I think of how successful Apple was during the years AFTER 2010.

He's worried because that success had ZERO to do with the Mac. It's all iPhone money. Apple hasn't done anything to the Mac in 10 years except strip out ports and make it thinner/less powerful than currently available PC hardware. So it is fairly understandable to be worried that Apple is potentially making this move.

 

[- rob -]

I appreciate any attention MR gives to the oncoming ARM-based mac, but I hope more work will go into thought pieces on this subject in the future.

The title is When Might Apple Release an Arm-based Mac, however only the final paragraph touches on this, but gives no other information about the forces that would move the ship date. Just links out to a prior entry that links out to a Bloomberg report. A more appropriate title would be "An introduction Apple's Expected Release of an Arm-based Mac"

As an overview on the topic it is missing insight and considerations that should have been included:

To me, the best candidate for an ARM-based Mac would be a new iteration of the 12" MacBook.

The article neglects to mention that Macbook, the center of the Macbook brand disappeared this past year! This is a big deal and should have been mentioned. The 12-inch, low-power Macbook but high price mac is the obvious first candidate for an ARM-based mac. There's a real business case that the Air--a favorite among corporate users could maintain the x86 architecture to avoid disrupting customer workflows with a new processor. Wheras a reintroduction of the Macbook with ARM and the restrictions that might place initially on software availability makes far more sense.

The entire article doesn't mention 'Swift' at all which is really surprising. It is not about 'will they make the transition' but when. They are putting all the little bits and pieces into place to make this transition happening and have been planning it for years. They already have all software vendors programme their software in Swift allowing them to release their software on all platforms at the same time.

If the title of the blog entry was followed in the content, this would have been a major component of "when." The article should reference major moves by Apple as long-term strategic changes, where the company introduces relatively minor but significant changes that lead to major change. (For example, ARKit in advance of an AR device)

Mentioning this ability to compile iOS apps for MacOS--and its announcement to developers first (Apple's most important customers of making major big changes succeed) would have been an opportunity to "connecting the dots looking forward" to use a Jobsian expression that I'd hope would have made it into a carefully researched and edited article on this topic from MacRumors.

 

[Oculus Mentis]

Worth reading to keep things in perspective...

Four years later: Why did Apple drop PowerPC?

Why did Apple drop the IBM-backed PowerPC architecture and switch to Intel in 2005? Four years to the month, one person involved in the discussions between Apple and IBM offers some insight.

www.cnet.com

 

[BOOMBA]

I am dumping Macs, because at the end of the day, Tim Cook's Apple no longer provides a computer with the horsepower I need - Sorry, but the 7,1 is $1,500 dollars of parts, which, outside of the CPU are EoL technology. A $4,500 Apple tax is just a bit too much.

Have fun with Candy Crush levels of software.

This is sadly where I am. I want the Mac Pro to be released soon so I can see some benchmarks. But barring something surprisingly amazing, I'll likely pull the trigger on a $3200 PC build I made on PC Part Picker that will kill the new Mac Pro (at least the new Mac Pro under $13,000)

 

[threesixty360]

I just wonder whether an arm based mac makes any sense in a market that is getting smaller and more niche.

I think we are more likely to see iPads morphing into the low cost computing line than MacBooks getting cheaper with arm chips.

what is actually the difference between an iPad Pro with a mouse and MacBook arm apart from the style of OS? In terms of power they will be the same just that the iPad will have tons more software available.

I think an iPad with a built in keyboard and track pad would be the game changer for rather than a MacBook running and. The later just presents more problems than solutions and the iPad w/keyboard almost creates a new market.

 

[obamtl]

One approach that’s not been discussed in this story is Apple going through a transition phase where an ARM processor is the primary CPU, with an Intel processor as the coprocessor. They’ve been paving the way for this on the software side, porting their apps over and making it easier to build ‘iOS’ apps that work on Mac.

 

[BOOMBA]

Eye for one feel that Apple has lost their heart. While changing processors won't do them 'arm, it could kneedlessly hit them in the hip pocket. Legs see what the future brings.

Who nose?

 

[andiwm2003]

a lot of people are very much in favor of arm based macs. i get the feeling a lot of them are casual users who also game. i wonder if they realize that it will be much harder to get games on the mac once they run on a proprietary niche platform compared to windows machines.

 

[Bustycat]

If they released right now, Apple would lose the professional market completely.

 

[andiwm2003]

Who nose?

I like to know. I'm all ears.