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

Page 13 - 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
4,956
7,674
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

Doug S

Platinum Member
Feb 8, 2020
2,269
3,522
136
Well, remember that this theoretical Apple SOC would be up against more advanced x86 CPUs than what are currently available. Supposedly the transition would start by next year, so that means Zen 3 for AMD and possibly Alder Lake for Intel.

No it won't be. It will be up against whatever x86 CPU is in the Macbook Air being sold at the time, and up against whatever x86 CPU is in any other Macs being sold when those go ARM.

There won't ever be a Mac sold with a Zen 3 installed, so its performance is irrelevant for whether an ARM Mac is faster than an x86 Mac. The fastest Macbook Air currently has an i5 able to boost up to 3.5 GHz. Surely no one claims that it is faster than the A13, or that the A14 will be slower than the A13. So obviously an ARM Macbook Air will easily beat the performance of an x86 Macbook Air.

How a hypothetical Macbook Air that has a Zen 3 or Alder Lake installed might perform is irrelevant, because we won't ever see such a beast if Apple is planning to release an ARM Macbook Air next year. Sure, there will be Windows laptops available with the Zen 3 etc. but there have always been Windows laptops available with faster CPUs than the Macbook Air. People aren't buying them to get the fastest laptop in the world.
 
  • Like
Reactions: name99

jpiniero

Lifer
Oct 1, 2010
14,629
5,247
136
Apple does need to release a dev kit of some sort. Perhaps it also double as an entry level laptop, and would come with an A13.

"Cough"

Are you sure about that? ;)

Since it looks more likely now they aren't using Renior, it seems unlikely they will use Ryzen any time soon.
 

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
Currently the low-end macbook air is dual core. So maybe it will go against a quad-core but still with double the cores. Even then I can't see a 8/4 lose to low end macbook air especially not in MT.

I think apple can scarifce some ST speed an still stay clearly ahead. We are talking mobile here. The base frequency of the intel i3 in low end macbook air is 1.1ghz!!! Apple could even reach clock parity at least in MT.

OK I looked at the rumors again concerning the timeframe, and it appears they will be going up against Tiger Lake CPUs with more mature stepping. It will probably be several years until they can come out with a desktop capable CPU that can compete with Intel and AMD.

Do browsersa and office really use SIMD? I doubt it...

When I made that comment on SIMD, I was thinking desktops, which is much further away for Apple. That said, there is some movement regarding the usage of 128 bit SIMD in Webassembly for certain algorithms.
 

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
No it won't be. It will be up against whatever x86 CPU is in the Macbook Air being sold at the time, and up against whatever x86 CPU is in any other Macs being sold when those go ARM.

When I made that comment I thought that Apple was targeting desktop capable CPUs for next year, but obviously that is incorrect. So if they release an SOC next year for laptops, it will be competing against Tiger Lake on a more mature stepping.

Tiger Lake is a strong core, with 20%+ average IPC improvements and AVX-512 support over Intel's current offerings. We'll have to wait till later this year to have a better idea of what Tiger Lake will be capable of and how a possible Apple laptop SOC would fare against it.
 

Ajay

Lifer
Jan 8, 2001
15,468
7,873
136
Cac is capacitance in the system seen during dynamic transitions. This includes the parasitic capacitance of the transistors (mainly gate capacitance) as well as the routing. So when you have a logic gate, the capacitance of the following gate (plus routing) is what you have to drive in order to actually pass your signal. Again, this capacitance is parasitic and induces power loss (leakage). Additionally, the faster you want to go, the more dynamic leakage you will have due to the frequency response of capacitive impedances. This is different from static leakage which is the leakage that occurs while the logic gate is in steady state or "holding" the signal.
Thanks. Going thru the math again via Wikipedia and other sources to get my brain cells working again. I majored in Physics and took some EE courses as well (wish I chose it as a minor) - so by physical, I mean mathematical :p
 

Glo.

Diamond Member
Apr 25, 2015
5,711
4,559
136
If you see a Mac with an AMD CPU you can be 100% sure you will never see one with an ARM CPU. They aren't going to switch horses from Intel to AMD immediately before switching from x86 to ARM.
As I said, the picture may be more complicated than you think it is.

ARM Macs purpose may be just entry level products. As I have said, x86 will have its place for the foreseeable future, and it will be HPC market. ARMs place may be the low-end/entry-level stack in computing world.

I do not see it likely that one possibility is exluding other, its more of a complimentary product stack rather than anything else.
 

moinmoin

Diamond Member
Jun 1, 2017
4,956
7,674
136

Doug S

Platinum Member
Feb 8, 2020
2,269
3,522
136
As I said, the picture may be more complicated than you think it is.

ARM Macs purpose may be just entry level products. As I have said, x86 will have its place for the foreseeable future, and it will be HPC market. ARMs place may be the low-end/entry-level stack in computing world.

I do not see it likely that one possibility is exluding other, its more of a complimentary product stack rather than anything else.


I don't buy that. It would be stupid to maintain both ARM & x86 Macs as a long term strategy. They get none of the benefit of designing every part of the Mac if they continued selling x86 Macs, and add the cost of supporting a second architecture long term. There would also be blowback from developers who wouldn't want to support two architectures going forward.
 

moinmoin

Diamond Member
Jun 1, 2017
4,956
7,674
136
@moinmoin

I think there are a lot of valid questions about how any CPU is going to perform under a full implementation of MacOS.
The base (non-graphical userland with XNU as kernel) of all Apple systems is Darwin, so that already exists for a potential ARM based system in Mac form factors.

The whole question is with what the graphical interface is being filled. There the two extremes are the proponents that macOS as is is there to stay, and on the other side that iOS (or more specifically all the adaptions in what's now called iPadOS) as is is already close to being usable in Mac form factors, and any potential mix between the two.

Performance is only a question once it's clear which audience is supposed to use which software on what system. I think too many people discussing this topic approach it as "hardcore" PC users that happen to use Mac systems, an audience Apple imo never intended to target directly but rather was covered as a side effect by switching to Intel. But it's also why I included AMD in this thread's topic. After all if the question indeed were of pure performance, Intel currently being superseded by AMD would need to lead to its chip appear at least in the high performance Mac products where Intel has no competitive counterparts.
 

moinmoin

Diamond Member
Jun 1, 2017
4,956
7,674
136
@DrMrLordX

I'd expect Darwin and all the ARM based user interfaces to be much more optimized on ARM at this point than OS X/macOS on Intel ever was (I consider that a given looking at their respective market size over the years).
 

moinmoin

Diamond Member
Jun 1, 2017
4,956
7,674
136
@DrMrLordX

My whole point was that Apple rather obviously seems to have no interest in optimizing for Intel hardware for some time now, neither macOS as UI nor Darwin as underlying OS and XNU as kernel. (As we know by now the scheduler has a big impact on performance so the most likely unoptimized piece of software would be XNU.)
 

Richie Rich

Senior member
Jul 28, 2019
470
229
76
One thing that I really hate. Having to depend on 14 year old benchmarks for ARM devices, and then having users tell us how valid they are.\

@Etain05 , SO 14 YEAR OLD BENCHMARKS ARE SUPPOSED TO BE GOOD ? THATS WHY YOU DOWNVOTED ME ?
And GeekBench 5.1 is outdated too? Because it shows same 80% IPC advantage of A13 over Zen2. So technically speaking SPECint2006 scales still pretty good and there is no reason to stop using it. Especially when you have a lot of measured data so you can compare uarchs among each other.

Regarding A14 predictions from Bloomberg article:
  • a conservative 10% of IPC increase would mean move from 1.84x PPC advantage to 1.1 x 1.84 = 2.02x .... so basically double Zen2's PPC.
  • 5nm allows 30% better power efficiency and 18% higher clock speed
  • assuming 10% IPC increase will consume 15% more power (half from that 30% 5nm advantage) A14 can use 15% power budget for clock increase, this is 1.15^1/3= 1.047 clock increase x 2.65 = 2.77 GHz.
  • overall performance could be 1.1 x 1.047 = 1.15 .... +15% over A13 withing same power envelope for iPhone 12
  • iPhone 12 - A14@2.77GHz performance equal to Zen2 at 5.6 GHz (2.77 x 2.02)
  • iPhone 11 - A13@2.65GHz was equal to Zen2@4.9 GHz (2.65 x 1.84)
  • but for MacBook they can use 18% higher clocks which means 2.65 GHz x 1.18 = 3.12 GHz
  • MacBook A14@3.12GHz would be equal to Zen2@6.3 GHz ... far more powerfull than any Renoir or Intel's cores
  • MacBook 8-core A14 TDP would be (A13's 3.5W x 1.64 from clock increase / 1.15 by 5nm eff = 4.99W per core x 8 = 40W total)... so no possible to run 8-cores at max clock
  • For 15W TDP it's (40/15)^(1/3) = 1.38x clock decrease -> 3.12GHz / 1.38 = 2.25 GHz for all-core load..... equal to Zen2@ 4.6 GHz

So the Bloomberg info is correct that Intel's 8-core CPUs will be faster. But only in desktop at 160W TDP at 5 GHz. In laptops it will have no competitor in native code. But if they consider emulation penalty this will will downgrade A14 bellow Intel's chips. However native code performance will be stellar.
 

Glo.

Diamond Member
Apr 25, 2015
5,711
4,559
136
So the Bloomberg info is correct that Intel's 8-core CPUs will be faster. But only in desktop at 160W TDP at 5 GHz. In laptops it will have no competitor in native code. But if they consider emulation penalty this will will downgrade A14 bellow Intel's chips. However native code performance will be stellar.
Reading comprehension, eh?

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.

So what is wrong? Apple themselves, or Workloads in Benchmarks are completely misleading, compared to real world?
 
  • Like
Reactions: Tlh97

Richie Rich

Senior member
Jul 28, 2019
470
229
76
Reading comprehension, eh?

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.

So what is wrong? Apple themselves, or Workloads in Benchmarks are completely misleading, compared to real world?
I can tell you what's wrong. You don't use any single number based on real measurements to support your opinion.

Real measured data are:
  • A13 has 1.84 times higher integer PPC than Zen2
  • A13@2.65GHz is equal to Zen2@4.9 GHz (2.65 x 1.84)

An old A13 in 8-core config would have devastating performance in laptops over any x86 chip. The only competitor within 15W TDP is Cortex-A77 however it would need 12-cores to keep up in MT load. Any x86 will loose either in ST or MT load in 15W TDP. That's clear from actual measured data.

Gurman said..... Richie Rich says that A14 will have an increased signal speed over speed of light resulting in power generation from heat extraction instead power burning. But Apple will never use it because iPhone would be uncomfortably cold. No numbers, no validity. Like in science or engineering.
 
Last edited:

Glo.

Diamond Member
Apr 25, 2015
5,711
4,559
136
I can tell you what's wrong. You don't use any single number based on real measurements to support your opinion.

Real measured data are:
  • A13 has 1.84 times higher integer PPC than Zen2
  • A13@2.65GHz is equal to Zen2@4.9 GHz (2.65 x 1.84)

An old A13 in 8-core config would have devastating performance in laptops over any x86 chip. The only competitor within 15W TDP is Cortex-A77 however it would need 12-cores to keep up in MT load. Any x86 will loose either in ST or MT load in 15W TDP. That's clear from actual measured data.

Gurman said..... Richie Rich says that A14 will have an increased signal speed over speed of light resulting in power generation from heat extraction instead power burning. But Apple will never use it because iPhone would be uncomfortably cold. No numbers, no validity. Like in science or engineering.
And you do? Where are Blender comaprisons, where is Final Cut Pro X, Logic Pro X compared to between x86 and Apple ARM? Where is code compilation? Nothing.

How the bloody hell do you dare to call Real measurements, a synthetic benchmark from 14 years ago?!

Where is anything that is REAL WORLD that we can compare to? Objectively!
 

Doug S

Platinum Member
Feb 8, 2020
2,269
3,522
136
And you do? Where are Blender comaprisons, where is Final Cut Pro X, Logic Pro X compared to between x86 and Apple ARM? Where is code compilation? Nothing.

How the bloody hell do you dare to call Real measurements, a synthetic benchmark from 14 years ago?!

Where is anything that is REAL WORLD that we can compare to? Objectively!


Well someone probably should do a comparison of stuff like Final Cut Pro since presumably Apple would do a good port. I remember someone did a comparison of Mathematica when that was ported to the iPad a few years ago, but while many things were similar in performance (this was several generation of Apple SoC ago too) certain tests were like 10x slower. He wasn't sure if that was because of a poor port or something that was multithreaded in the x86 version but not in the iPad version. It has been a few years, maybe the iPad port of Mathematica is better now.

As for compilation there are two sources you can use, SPEC2006 includes a gcc test and Geekbench includes an LLVM test. Both show Apple's ARM easily outperforming x86, so I guess you will say those tests don't count and you want something you would consider a "real" compilation test (i.e. something that will show the x86 outperforming Apple)
 

Glo.

Diamond Member
Apr 25, 2015
5,711
4,559
136
Well someone probably should do a comparison of stuff like Final Cut Pro since presumably Apple would do a good port. I remember someone did a comparison of Mathematica when that was ported to the iPad a few years ago, but while many things were similar in performance (this was several generation of Apple SoC ago too) certain tests were like 10x slower. He wasn't sure if that was because of a poor port or something that was multithreaded in the x86 version but not in the iPad version. It has been a few years, maybe the iPad port of Mathematica is better now.

As for compilation there are two sources you can use, SPEC2006 includes a gcc test and Geekbench includes an LLVM test. Both show Apple's ARM easily outperforming x86, so I guess you will say those tests don't count and you want something you would consider a "real" compilation test (i.e. something that will show the x86 outperforming Apple)
How large is the code that has to be compiled in those tests? How does it load the cores?

I want to see code compilation done by benchmarkers, not the automated benchmark the emulates real world workloads. That is real comparison.