Solved! ARM Apple High-End CPU - Intel replacement

Page 23 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Richie Rich

Senior member
Jul 28, 2019
470
229
76
There is a first rumor about Intel replacement in Apple products:
  • ARM based high-end CPU
  • 8 cores, no SMT
  • IPC +30% over Cortex A77
  • desktop performance (Core i7/Ryzen R7) with much lower power consumption
  • introduction with new gen MacBook Air in mid 2020 (considering also MacBook PRO and iMac)
  • massive AI accelerator

Source Coreteks:
 
  • Like
Reactions: vspalanki
Solution
What an understatement :D And it looks like it doesn't want to die. Yet.


Yes, A13 is competitive against Intel chips but the emulation tax is about 2x. So given that A13 ~= Intel, for emulated x86 programs you'd get half the speed of an equivalent x86 machine. This is one of the reasons they haven't yet switched.

Another reason is that it would prevent the use of Windows on their machines, something some say is very important.

The level of ignorance in this thread would be shocking if it weren't depressing.
Let's state some basics:

(a) History. Apple has never let backward compatibility limit what they do. They are not Intel, they are not Windows. They don't sell perpetual compatibility as a feature. Christ, the big...

insertcarehere

Senior member
Jan 17, 2013
639
607
136
Zen cores could, in theory, run in the same power envelope as a high end Snapdragon while providing better performance.

If Zen cores could actually do such a thing, we'd be seeing AMD SoCs in phones.

At least, that's the logic some posters here are displaying with regard to ARM on MacBooks.
 

Doug S

Platinum Member
Feb 8, 2020
2,202
3,405
136
Thanks for great answer!
OK, so we can expect A14X right after A14.
Apple can announce MacBook Air on ARM this year in theory. I think Apple can have sell both Intel and ARM versions of laptops in the same time. This is different situation nowadays than move from PowerPC because that time was PPC slower than Intel so laptops with Wintel were performing much better. This time Apple can ask premium price for premium A14X performance and battery life.

They can't announce an ARM Mac without giving developers time to port some applications so it launches with plenty of native titles ready to announce. Same thing they did for the last couple transitions.
 

Doug S

Platinum Member
Feb 8, 2020
2,202
3,405
136
I don't understand why people have said that Apple's ARM CPUs are faster than x86. I have, in my possession, an iPhone Pro Max. I have run various benchmarks on it, and I've also looked up Benchmarks online. The CPU is still very much in line with a Core i3 or low end i5 when it comes to performance.


So tell us, which "benchmarks" have you run that show it performing like an i3? I trust Anandtech's crew to run benchmarks more than some rando on a forum who won't even name what benchmarks he claims he's using.
 
  • Like
Reactions: Etain05

Richie Rich

Senior member
Jul 28, 2019
470
229
76
So tell us, which "benchmarks" have you run that show it performing like an i3? I trust Anandtech's crew to run benchmarks more than some rando on a forum who won't even name what benchmarks he claims he's using.
He talks about Geekbench5 score in MT which is true:

MT score is not fully comparable because A13 is only 2 big core + 4 little core and Intels are 4 big cores + SMT.
You would need to compare with 4 big core A12X from iPad Pro GB5 score. And bang, an old A12X has MT score of 4607 and outperforming any 4-core x86 laptop chip by huge margin (while running at just 2.5 GHz).


If we isolate just single core performance than results are very different and suddenly Apple A13 is outperforming Intel by large number:

And Apple A13 is running within 5W TDP, A12X 7W TDP...….. in compare to Intel laptops with 15W TDP and ST core running over 4 GHz (and still loosing in every way).
There is no point to answer some others - haters never write numbers and facts.

EDIT: Added Cortex A77 in new Snapdragon into comparison. It's 4+4 big.LITTLE and is on par with 4-core x86 Intels in MT load (with much lower power consumption). ST load is lower due to max clock at 2.8 GHz. IMHO biggest danger for Intel and AMD will come from generic Cortex cores such as this A77 or new A78. Don't forget that A77 has slightly higher IPC/PPC than Zen2 (+8% according to SPECint2006). This is huge milestone for generic Cortex core having higher IPC/PPC for the first time in history. Also very interesting from uarch point of view - A77 has 4xALU + 2xjump units, this half way to Apple's 6xALU design when Apple's core uses 4xALU + 2x simpleALU/jump units (and clearly wider than x86 stuck at 4xALU for decade). I can speculate that A78 might upgrade those 2xjump units into 2x simpleALU/jump in Apple style (seems logical and evolutional step to me but who knows). When every cheap Raspberry PI will have A78 with IPC/PPC higher than Ice Lake and Zen3, this is the biggest threat for x86 laptops and desktops. Apple is great technological demonstrator for future ARM Cortex cores though.
 
Last edited:
  • Like
Reactions: Etain05

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,480
14,434
136
That is a bit naive in a day and age customers still buy intel xeons over AMD Epyc. If only raw performance would matter, intel should make close to 0 server cpu sales.
Well, at least Google, Microsoft, cloudflare and myself have the brains to buy EPYC.

No accounting for taste or brains for the others....
 
  • Like
Reactions: Tlh97

soresu

Platinum Member
Dec 19, 2014
2,617
1,812
136
When every cheap Raspberry PI will have A78 with IPC/PPC higher than Ice Lake and Zen3
Don't hold your breath, they only just moved to 28nm A72 with RPi 4, and they probably won't upgrade to A78 even when they do make a change.

The RPi guys are going with the cheapest possible process to fab their chips on, so at best they might make the switch to 8nm sometime after 5nm/3nm is dropped by high end customers, but only if 8nm becomes cheap enough to make it worthwhile for them.
 

soresu

Platinum Member
Dec 19, 2014
2,617
1,812
136
If only raw performance would matter, intel should make close to 0 server cpu sales.
Raw perf probably does matter to server/datacenter buyers - but likely not nearly as much as perf/watt - cost of running both the chips AND the cooling is a big consideration over and above the performance itself.
 

jpiniero

Lifer
Oct 1, 2010
14,510
5,159
136
Well, at least Google, Microsoft, cloudflare and myself have the brains to buy EPYC.

No accounting for taste or brains for the others....

AWS did announce they were going to have Rome instances but never followed through. They do have first gen Epyc instances available. Not sure what happened, probably a combination of Intel offering a very nice deal and wanting to push Gravitron.
 

Richie Rich

Senior member
Jul 28, 2019
470
229
76
Don't hold your breath, they only just moved to 28nm A72 with RPi 4, and they probably won't upgrade to A78 even when they do make a change.
RPi 6 could have A78 :)
RPi 5 could be realistically something 14nm A76 based in 2021/22. Even A76 is huge jump from slow A72.
 

Thunder 57

Platinum Member
Aug 19, 2007
2,647
3,706
136
Sigh... and here I thought we were done with this. The problem is as long as there is no product to compare, this thread can go on indefinitely. There's just no proof either way.

The fact that there is no product leads me to believe your wild speculation simply doesn't hold up. Apple aren't a bunch of idiots. If they could relase a laptop that would blow away AMD/Intel while using much less power, it would've happened by now. Maybe it's in the works now, but until we start to see some numbers, there is no point in continuing this speculation.

I mean first of all you are assuming A12/A13/A14 can clock comparatively. That's a huge assumption. AMD/Intel mobile CPU's still boost to 4GHz+, power/thermal permitting. Surely Apple's chip would boost higher if they could.
 
  • Like
Reactions: USER8000

JasonLD

Senior member
Aug 22, 2017
485
445
136
Even though we are seeing the reports of Apple moving to their own processors for their Macs (like every year for 3-4 years?). Apple probably isn't in hurry to make the move anytime soon. At least not until they have architecture ready to make a sweeping change within a year. (PowerPC to Intel transition didn't take more than a year for every Mac product).
Going by SPECint, it does look like they have really capable architecture, but I have yet to see any indication that it can scale up to many cores that would be suitable for high end iMacs and Mac Pros. Maybe we might see the glimpse of it when they announce A14 architecture, if not, then we probably won't see the transition until 2022 at the earliest.
 
  • Like
Reactions: beginner99

Doug S

Platinum Member
Feb 8, 2020
2,202
3,405
136
I mean first of all you are assuming A12/A13/A14 can clock comparatively. That's a huge assumption. AMD/Intel mobile CPU's still boost to 4GHz+, power/thermal permitting. Surely Apple's chip would boost higher if they could.

Obviously you haven't been following the thread, or seen Anandtech's benchmarking of Apple's SoCs. They don't need to clock at 4 GHz or even 3 GHz to be competitive with Intel's high clocked CPUs, the one that's in the iPhone 11 is competitive with them!
 

Thunder 57

Platinum Member
Aug 19, 2007
2,647
3,706
136
Obviously you haven't been following the thread, or seen Anandtech's benchmarking of Apple's SoCs. They don't need to clock at 4 GHz or even 3 GHz to be competitive with Intel's high clocked CPUs, the one that's in the iPhone 11 is competitive with them!

Still not buying it. Let's see one do real work. As Mark (I think) said, why aren't Amazon/MS/etc running a bunch of phone chips in their servers? There's far more to it.

As for clocking higher, I was going by Richie Rich's own (optimistic, IMO) numbers.

1. Intel Core i9 9900K @5GHz ......... SPECint2006 score: 54.28 ...... 10.86 pts/GHz
2. Apple A13 @2.65 GHz .................. SPECint2006 score: 52.82 ...... 19.93 pts/GHz ...... +83 % IPC over 9900K
3. AMD Ryzen 3950X @4.6 GHz ...... SPECint2006 score:50.02 ...... 10.87 pts/GHz ...... + 0% IPC over 9900K .... fastest clocked Ryzen beaten by iPhone CPU
4. ARM Cortex A77@2.84 GHz ......... SPECint2006 score: 33.32 ...... 11.73 pts/GHz ...... + 8% IPC over 9900K


So Apple's A13 would need to clock considerably higher for it to take the massive lead he thinks it's capable of. If you're not going to clock it higher, what's the point? You'll get similar performance and lose compatibility. So no, don't tell me I "obviously" haven't been paying attention.
 
  • Like
Reactions: Tlh97 and USER8000

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
If Zen cores could actually do such a thing, we'd be seeing AMD SoCs in phones.

At least, that's the logic some posters here are displaying with regard to ARM on MacBooks.

AMD has 14nm parts with a 6 watt TDP. They also have Renoir with a 15 watt TDP. It stands to reason that a 7nm part could easily operate in half the TDP. Of course, AMD won’t do this because there is no demand.

As far as other comments about Apple chips being “faster”, I will believe it when I see it. Running benchmarks is one thing. Doing development work or video/graphics work is something else. They would need to drastically increase the core count, which would negate any power savings.
 

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
He talks about Geekbench5 score in MT which is true:

MT score is not fully comparable because A13 is only 2 big core + 4 little core and Intels are 4 big cores + SMT.
You would need to compare with 4 big core A12X from iPad Pro GB5 score. And bang, an old A12X has MT score of 4607 and outperforming any 4-core x86 laptop chip by huge margin (while running at just 2.5 GHz).


If we isolate just single core performance than results are very different and suddenly Apple A13 is outperforming Intel by large number:

And Apple A13 is running within 5W TDP, A12X 7W TDP...….. in compare to Intel laptops with 15W TDP and ST core running over 4 GHz (and still loosing in every way).
There is no point to answer some others - haters never write numbers and facts.

EDIT: Added Cortex A77 in new Snapdragon into comparison. It's 4+4 big.LITTLE and is on par with 4-core x86 Intels in MT load (with much lower power consumption). ST load is lower due to max clock at 2.8 GHz. IMHO biggest danger for Intel and AMD will come from generic Cortex cores such as this A77 or new A78. Don't forget that A77 has slightly higher IPC/PPC than Zen2 (+8% according to SPECint2006). This is huge milestone for generic Cortex core having higher IPC/PPC for the first time in history. Also very interesting from uarch point of view - A77 has 4xALU + 2xjump units, this half way to Apple's 6xALU design when Apple's core uses 4xALU + 2x simpleALU/jump units (and clearly wider than x86 stuck at 4xALU for decade). I can speculate that A78 might upgrade those 2xjump units into 2x simpleALU/jump in Apple style (seems logical and evolutional step to me but who knows). When every cheap Raspberry PI will have A78 with IPC/PPC higher than Ice Lake and Zen3, this is the biggest threat for x86 laptops and desktops. Apple is great technological demonstrator for future ARM Cortex cores though.

It is important to note that a Ryzen 3700X (65W TDP) gets around the same SC score as the Apple chip, however, Geekbench has already been shown to favor Apple’s desktop operating system, so we don’t know what optimizations that Apple has under the hood. Chip performance could very likely fall apart due to the “open” nature of the Mac. When you combine this with the fact that Apple would have to use an emulator, the prospects of having better performance go out the window.

I don’t use Intel chips here, because they only have very specific parts on 10nm.
 

insertcarehere

Senior member
Jan 17, 2013
639
607
136
AMD has 14nm parts with a 6 watt TDP. They also have Renoir with a 15 watt TDP. It stands to reason that a 7nm part could easily operate in half the TDP. Of course, AMD won’t do this because there is no demand.

There is no demand for a x86 Smartphone SoC (which you believe AMD can build) which can match/beat its ARM conpetitors in performance? Qualcomm has over 3x AMD's revenues, the majority of which come from smartphone chips, demand is not the issue here.
 
  • Like
Reactions: Etain05

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
There is no demand for a x86 Smartphone SoC (which you believe AMD can build) which can match/beat its ARM conpetitors in performance? Qualcomm has over 3x AMD's revenues, the majority of which come from smartphone chips, demand is not the issue here.

It’s less to do with demand and more to do with margins. Android runs fine on x86, and I suspect iOS would as well. However, Intel and AMD build big, fast, high margin chips.

I read the Anandtech article again and it baffles me why people think that this chip is ready for the desktop. The only thing it has going for it is integer performance. It’s memory subsystem is horrible, and floating point operations are also much slower than x86.
 

insertcarehere

Senior member
Jan 17, 2013
639
607
136
I read the Anandtech article again and it baffles me why people think that this chip is ready for the desktop. The only thing it has going for it is integer performance. It’s memory subsystem is horrible, and floating point operations are also much slower than x86.

It baffles me that you believe a hypothetical Apple ARM laptop/desktop chip couldn't possibly improve its memory subsystem to be better than one currently designed for phone/tablet.

Likewise, losing by ~15% in FP to the very fastest single-threaded performance x86 chips on offer (certainly faster than equivalent gen laptops would be) in a device form factor barely bigger than a cpu socket strikes me as a lot of things, but "non-competitive" is not it.
 

Thala

Golden Member
Nov 12, 2014
1,355
653
136
As far as other comments about Apple chips being “faster”, I will believe it when I see it. Running benchmarks is one thing. Doing development work or video/graphics work is something else. They would need to drastically increase the core count, which would negate any power savings.

Efficiency does not decrease when you have more cores. And of course an hypothetical 15W TDP A-chip would have more than 4 cores with increased frequency putting anything from AMD to shame in the same power envelope.
There are no numbers indicating that AMD even plays in the same league.
 

Doug S

Platinum Member
Feb 8, 2020
2,202
3,405
136
So Apple's A13 would need to clock considerably higher for it to take the massive lead he thinks it's capable of. If you're not going to clock it higher, what's the point? You'll get similar performance and lose compatibility. So no, don't tell me I "obviously" haven't been paying attention.


Apple doesn't need a "massive lead", or any lead at all. They just need ARM based Macs to not be a step down from x86 based Macs, and they have reached that point. Why should they care about x86 compatibility, they only need that during the transition. They don't support running 68K or PPC code any longer, in a decade they wouldn't support x86 code any longer either.

Apple wants control over the Mac platform like they have over the iPhone platform. Today they have to work according to Intel's schedule and priorities, and when Intel slips like with their ongoing 10nm fiasco Apple is forced to change their plans.