Discussion Quo vadis Apple Macs - Intel, AMD and/or ARM CPUs? ARM it is!

Page 15 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,618
136
Due to popular demand I thought somebody should start a proper thread on this pervasive topic. So why not do it myself? ;)

For nearly a decade now Apple has treated their line of Mac laptops, AIOs and Pro workstations more of a stepchild. Their iOS line of products have surpassed it in market size and profit. Their dedicated Mac hardware group was dissolved. Hardware and software updates has been lackluster.

But for Intel Apple clearly is still a major customer, still offering custom chips not to be had outside of Apple products. Clearly Intel is eager to at all costs keep Apple as a major showcase customer.

On the high end of performance Apple's few efforts to create technological impressive products using Intel parts increasingly fall flat. The 3rd gen of MacPros going up to 28 cores could have wowed the audience in earlier years, but when launched in 2019 it already faced 32 core Threadripper/Epyc parts, with 64 core updates of them already on the horizon. A similar fate appears to be coming for the laptops as well, with Ryzen Mobile 4000 besting comparable Intel solutions across the board, with run of the mill OEMs bound to surpass Apple products in battery life. A switch to AMD shouldn't even be a big step considering Apple already has a close work relationship with them, sourcing custom GPUs from them like they do with CPUs from Intel.

On the low end Apple is pushing iPadOS into becoming a workable mutitasking system, with decent keyboard and, most recently, mouse support. Considering the much bigger audience familiar with the iOS mobile interface and App Store, it may make sense to eventually offer a laptop form factor using the already tweaked iPadOS.

By the look of all things Apple Mac products are due to continue stagnating. But just like for Intel, the status quo for Mac products feels increasingly untenable.
 
  • Like
Reactions: Vattila

Richie Rich

Senior member
Jul 28, 2019
470
229
76
As for the difference between CL version of Blender versus GUI, until we can prove that there's no difference, I don't want to assume anything. That's all. Blender may not see a big hit (I don't know that for sure), but I do think that the majority of consumer facing programs rely heavily on the GUI, and removing it from the performance equation makes the benchmark far less useful.
How much CPU time can cost small picture tile update lets say every second? CPU does at 4GHz 4 bilion cycles at rendering code and then needs to spend few cycles on screen update which is mostly done by GPU anyway... this influence is like 0.001%
 

dacostafilipe

Senior member
Oct 10, 2013
771
244
116
What benchmark could we use to test those devices better?

If there's no app available, could we make a new one? What kind of code should we run?

We have a lot of iOS devices in the office and I could bring them home for testing if needed. I didn't touch Xcode for 1-2 years, but I could look into building an app when code available to test, even looking into porting code if needed.

If it's no to big of a project, I could invest some time into this.
 
  • Like
Reactions: lightmanek

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
We have a lot of iOS devices in the office and I could bring them home for testing if needed. I didn't touch Xcode for 1-2 years, but I could look into building an app when code available to test, even looking into porting code if needed.

If you really want to do it, I would start by looking at any command-line open-source application that suits your fancy. Phoronix test suite has a lot of that stuff. Blender's an obvious one (it does have a CLI version). But as Thala warned, there will be some porting involved, even if you avoid GUI applications.

Another option is to try using WASM. But that might introduce other problems. SIMD support under WASM is a work-in-progress.
 

soresu

Platinum Member
Dec 19, 2014
2,612
1,811
136
Another option is to try using WASM. But that might introduce other problems. SIMD support under WASM is a work-in-progress.
Waaaat? SIMD in asm.js was working years ago and they still haven't stood up WASM SIMD?

I get that they probably had to go back to the drawing board after Spectre/Meltdown, but this is ridiculous - same thing with the WebAudio standard, still not ratified to v1 after who knows how many years of work on it.
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Waaaat? SIMD in asm.js was working years ago and they still haven't stood up WASM SIMD?

This document is one of the most recent ones I've seen related to SIMD support in WASM:


Granted, that's particular to Google's implementation, but still.

And yes, I guess that means one could just ignore WASM and go for asm.js instead.
 

Gideon

Golden Member
Nov 27, 2007
1,608
3,564
136
This document is one of the most recent ones I've seen related to SIMD support in WASM:


Granted, that's particular to Google's implementation, but still.

And yes, I guess that means one could just ignore WASM and go for asm.js instead.

Well unfortunately asm.js is pretty much deprecated in favor of WASM :p

WASM will support for SIMD will come soon enough for both Cranelift (Firefox) and v8 (Chrome) I'm pretty sure it's there already in nightlies in some form.
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
Apple dropped a more powerful Macbook Pro 13" w/ a 10th gen Intel chip today

Doesn't really have as much forward-looking bearing, because even if Apple were planning on moving Pro to ARM next year, these MBP laptops have been getting yearly updates anyway, so there's nothing we could really glean from the release.

However, I do think it's interesting that Apple continue to offer a max of 4-core chips in their top tier laptop lineup (assuming the top line chip is the 4-core i7-1068NG7). In that case, Apple certainly aren't setting the bar terribly high for beating that performance with a possible move to ARM even on MBP lines.
 
  • Like
Reactions: lightmanek

Glo.

Diamond Member
Apr 25, 2015
5,657
4,407
136
Apple dropped a more powerful Macbook Pro 13" w/ a 10th gen Intel chip today

Doesn't really have as much forward-looking bearing, because even if Apple were planning on moving Pro to ARM next year, these MBP laptops have been getting yearly updates anyway, so there's nothing we could really glean from the release.

However, I do think it's interesting that Apple continue to offer a max of 4-core chips in their top tier laptop lineup (assuming the top line chip is the 4-core i7-1068NG7). In that case, Apple certainly aren't setting the bar terribly high for beating that performance with a possible move to ARM even on MBP lines.
Or paints a simple picture, that even 8/4 SoC ARM from Apple is not able to compete with 4C/8T x86 from Intel, let alone 8C/16T from AMD.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
Apple dropped a more powerful Macbook Pro 13" w/ a 10th gen Intel chip today

Doesn't really have as much forward-looking bearing, because even if Apple were planning on moving Pro to ARM next year, these MBP laptops have been getting yearly updates anyway, so there's nothing we could really glean from the release.

However, I do think it's interesting that Apple continue to offer a max of 4-core chips in their top tier laptop lineup (assuming the top line chip is the 4-core i7-1068NG7). In that case, Apple certainly aren't setting the bar terribly high for beating that performance with a possible move to ARM even on MBP lines.

Only the upgraded 13" has Ice Lake... the base model has Coffee Lake still.

Or paints a simple picture, that even 8/4 SoC ARM from Apple is not able to compete with 4C/8T x86 from Intel, let alone 8C/16T from AMD.

If they were going to stay with x86, I think they might have actually gone with Renior.
 

Glo.

Diamond Member
Apr 25, 2015
5,657
4,407
136
If they were going to stay with x86, I think they might have actually gone with Renior.
There is one thing that for now completely kills the idea of ARM/AMD MacBooks.

Thunderbolt 3. For MacBook they can release it without TB3, and then one year later they can obsolete it with adding USB4 which will have a TB3 protocol. But for MBPs, and Airs for now it is impossible for it to have anything else than Intel, unless Apple will want to pay more for licensing the TB3.
 

Doug S

Platinum Member
Feb 8, 2020
2,201
3,405
136
There is one thing that for now completely kills the idea of ARM/AMD MacBooks.

Thunderbolt 3. For MacBook they can release it without TB3, and then one year later they can obsolete it with adding USB4 which will have a TB3 protocol. But for MBPs, and Airs for now it is impossible for it to have anything else than Intel, unless Apple will want to pay more for licensing the TB3.


Intel made TB3 licensing royalty free three years ago, so that's not a roadblock at all. Plus it was originally developed in partnership with Apple, so Apple probably already had a royalty free license in perpetuity unless their lawyers were asleep when they made that agreement.
 

repoman27

Senior member
Dec 17, 2018
342
488
136
Apple has plenty of experience when it comes to these types of transitions, so it makes sense to look at the major ones that they've already pulled off for the Mac and what technologies they employed to make them happen:

680x0 > PowerPC
software emulation: Mac 68k emulator
executable format: fat binary

PowerPC 32-bit > 64-bit
executable format: Multi-Architecture Binary

Classic Mac OS > Mac OS X
transitional API: Carbon
virtualization layer: Classic Environment

PowerPC > x86
software emulation: Rosetta
executable format: Universal binary
native OS support: Boot Camp

x86-32 > x86-64
executable format: Universal binary

x86 > ARM
???

Whether it takes months or years for Apple to replace every x86 Mac in their lineup with an ARM version, the transition itself will take considerably longer. They need to make this period as painless as possible for both developers and end users alike, or they risk seriously damaging the platform. Even after the last x86 Mac is sold, Apple will still need to continue to support macOS on x86 for another ~5 years, as will many developers. Furthermore, many developers will take their sweet time or never even bother to build ARM versions of their Mac applications. Yet on day one that ARM Macs are available, near as makes no difference 100% of existing Mac software needs to "just work".

The only way to solve this problem is the tried-and-true combination of multi-architectural binaries and emulation. There were a several posts earlier in this thread maligning Rosetta, but the fact of the matter is it worked brilliantly. Yes, there was a distinct performance penalty, but without Rosetta the PowerPC to x86 transition would have been a non-starter. And of course some people still clenched when it was removed from Mac OS X Lion five and a half years later.

If we look at the numerous 32 to 64-bit transitions that have happened, we see a similar use of fat binaries, but no software emulation and subsequent performance penalty. This is because when 64-bit extensions are added to an existing architecture, affordances for backwards compatibility are included in the hardware. These are often complete enough that you can continue to run an unmodified 32-bit OS on 64-bit hardware as long you don't mind giving up the benefits of moving to 64-bit.

Apple is in a unique position here where they aren't just transitioning between two disparate architectures; they have complete control over the silicon. The obvious thing to do in such a situation is to include fixed function blocks in hardware to accelerate x86 emulation.

With the transition to x86, Apple introduced another piece that will affect them going forward. The x86 Mac platform enables less costly virtualization (rather than emulation) when hosting popular guest OSes, and with Boot Camp Apple even provides first-party support for running third-party OSes natively on Mac hardware. Now of course there are AArch64 versions of Linux and Windows 10 that presumably could run natively on ARM Macs under a system similar to Boot Camp. However, Microsoft's use of the WOW64 layer to provide software emulation for x86 on ARM is limited to 32-bit applications. It's also not fast. Apple would have a leg up here seeing as they already have faster ARM SoCs than anything we've seen Windows 10 running on thus far. However, if Apple's hardware accelerated emulation of x86 is solid enough, running x86-64 Windows as a guest OS may end up being the better option. I rather doubt that Apple would go so far as to include full, transparent x86 emulation in hardware, but there is always that possibility as well.

With the release of macOS 10.15 Catalina last fall, Apple finally removed support for Carbon and, by extension, 32-bit apps. At one point Universal binaries contained as many as four different versions targeting the 32 and 64-bit versions of both the PowerPC and x86 architectures. But now, the decks have been cleared. Apple is ready to make yet another transition.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136

Fresh rumor says the first ARM Macs will be the 13" MBP and an iMac at the end of the year or early next year.

Before then, the iMac is going to be updated very soon, presumably with Comet Lake, which will be the go to machine for people who can't migrate that quickly.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
Gurman specifically said, that 8/4 Core config will land in entry level products, below anything x86, because those ARM CPUs cannot tackle x86 in performance.

Glo, I read two of the articles and they mention it can't be in high end Macbook Pro, iMacs and Mac Pro. But in smaller laptops it'll be a different story.

The most recent article says especially in graphics and artificial intelligence.
 

coercitiv

Diamond Member
Jan 24, 2014
6,151
11,670
136
Even The Verge joined the choir. This will be lots of fun:
The analyst expects the new ARM machines to outperform their Intel predecessors by 50 to 100 percent, though the actual performance will of course depend on what Apple decides to prioritize with the new designs.
 
Last edited:

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136
That's a lot earlier than expected.

We will see. It's a rumor but yeah in some way wouldn't surprise me as the can use the same SOC as in ipad pro for the lower end devices.

Even The Verge joined the choir. This will to lots of fun:

Outperfomr 50-100 in what? Battery life? Maybe. Pure performance? Maybe in browser tests but I wonder how much all the developers using macbooks will like their compile times on these ARM chips. I think they are great for consumer devices (browsing performance, battery life) but not so much for content creation (anything from image processing to actual programming). Due to that apple must ask themselves if it's good that people making actual content will move out of their ecosystem? So either they will need to keep some x86 based devices or come up with a ARM core that also works with compilation (unless that is a graviton2 issue only).
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
The 13" MacBook Pro receiving an ARM chip will be interesting. Apple must have done some demographic evaluation and figured out that most of the 13" MBP users aren't actual Pro users -- or Apple have rapidly expedited full versions of the pertinent Pro software for these ARM laptops.

I still have my reservations about heat / performance of ARM in these laptops doing long runs of multicore processing, considering my iPhone 11 gets hot just running Zoom calls over cellular.
 

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136
The 13" MacBook Pro receiving an ARM chip will be interesting. Apple must have done some demographic evaluation and figured out that most of the 13" MBP users aren't actual Pro users -- or Apple have rapidly expedited full versions of the pertinent Pro software for these ARM laptops.

I still have my reservations about heat / performance of ARM in these laptops doing long runs of multicore processing, considering my iPhone 11 gets hot just running Zoom calls over cellular.

Or, they realised that in order to build an ARM ecosystem, they need developers running on ARM- so they prioritised the laptop that developers use.
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
Or, they realised that in order to build an ARM ecosystem, they need developers running on ARM- so they prioritised the laptop that developers use.
It seems like you're implying something different from what I said, which was "Apple have rapidly expedited full versions of the pertinent Pro software for these ARM laptops", though I'm not sure if that's what you meant.

It sounds like you're implying that Apple are releasing 13" MBP on ARM so that developers have something to develop on, implying that the Pro software users rely on that is now on x86 is not yet fully-developed for ARM? In that case, it means they will be forcing customers who may have purchased an MBP 13" into a higher end MBP that still runs on x86.

The whole thing seems like it could be a mess / fractured. Developers have to maintain both ARM and x86 versions of their pro software for Mac Pro/MBP x86, MBP ARM, iPad ARM, just on the Apple side.
 
Last edited:

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
The whole thing seems like it could be a mess / fractured. Developers have to maintain both ARM and x86 versions of their pro software for Mac Pro/MBP x86, MBP ARM, iPad ARM, just on the Apple side.

That's why Apple will encourage developers to use Apple's APIs, which are arch neutral.