Solved! ARM Apple High-End CPU - Intel replacement

[Richie Rich]

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:

 

[Richie Rich]

And 12 ALU's !!

It's not bad idea. 12xALUs with SMT2 would be smart alternative to current Apple A13 6xALU with no SMT.

 

[Steelbom]

They have done it before with Rosetta/POWER/x86, though I think it more likely that they would simply encourage software suppliers to make the jump long before they do so with hardware.

I believe that iPad OS is the beginning of this strategy for them, given its push towards a more laptop like feature set (keyboard/mouse/memory cards/USB C).

As to AMD, Apple have still yet to make a Mac using an AMD CPU as far as I am aware - even with Surface dipping its toes in to those waters I would not expect Apple to do the same until Intel are truly floundering, considering they were not even top dog when Apple made the jump to x86 in the first place.

They did but there were far fewer apps available for Mac back in those days, were there not? There's a lot more now, and a lot of big apps like Adobe that would be difficult to bring across.

Not to mention they need to maintain it for x86 and then ARM as well.

Imo if anything, Apple is more likely to build a x86 CPU then they are to use an ARM CPU as a Mac's processor.

I could see them partnering with someone like AMD to build a custom chip that has both ARM and x86 cores on it and using that during a transition period. Obviously Apple could run the OS and own applications on ARM because they have control over the code and could rewrite it. Legacy applications could be run on x86 cores as developers transition off of them over several years. Eventually the x86 cores just get replaced with emulation/translation for anything that won't get a modern/updated release.

That's probably possible, but at that point wouldn't they just look at making their own x86 CPU? Requiring all apps to support ARM would stagnate the app ecosystem, not that it's particularly huge compared to Windows. Developers would need to support both Mac + Mac ARM for years and years.

It is only a matter of time. Apple did this during the 68K to PPC transition and the PPC to X86 transition. Both times flawlessly.

Yeah but there were far fewer apps back then, and less big ones like Adobe, no? There's a lot of apps for Mac now, and quite a few big ones... having developers support both Mac + Mac ARM means more expense to them, and more time and effort.

 

[thunng8]

yes, fewer app back then, but Adobe apps were available for mac (in fact Adobe applications were on mac before windows) , so big applications were there including Microsoft Office..

 

[Steelbom]

yes, fewer app back then, but Adobe apps were available for mac (in fact Adobe applications were on mac before windows) , so big applications were there including Microsoft Office..

Interesting, never knew that!

 

[Thunder 57]

It's not bad idea. 12xALUs with SMT2 would be smart alternative to current Apple A13 6xALU with no SMT.

How has Apple's talent acquisition folks not found you yet??

 

[Bouowmx]

YouTube (technology commentary/story-time) is so cringe.

And of course, this twat has Patreon and merch.

 

[soresu]

yes, fewer app back then, but Adobe apps were available for mac (in fact Adobe applications were on mac before windows) , so big applications were there including Microsoft Office..

Interesting, never knew that!

Some of the current Mac apps like Serif Affinity Photo even have ARM versions already (for iOS anyway).

Adobe's efforts to port their Mercury engine to Vulkan compute allows execution on both Mac and iOS through MoltenVK - this has already born fruit in the Premiere Rush app for ARM platforms using the same compute back end, regardless of the UI code or whatever else Rush runs on.

Full Photoshop was also in the works for iOS last I checked, not sure if it's out yet though.

Obviously Office has apps just about everywhere.

I'd be extremely surprised if Apple has not had back room talks with many major software developers including Adobe, Autodesk, Steinberg (music/sound mixing) and Avid.

In much the same way as any computer hardware designer has to look to the future, it is inevitable that software companies can and must do the same in order to stay relevant in an evolving market.

Whatever the currently announced state of each major x86 software package is, I would be extremely surprised if there were not at the very least plans for a move/duplication of the codebase back end to ARMv8+ going forward.

 

[Failnaught]

What does Apple gain by going to ARM? Longer battery life? More control over the logistics/suppliers? How much of a performance or design advantage could they get with ARM, scaled up to desktop/laptop performance and complexity levels? Is it worth the (massive) pain on the software ecosystem front?

I suspect this is somewhat analogous to Huawei being forced to design their own future SOC for their phones post Trump. The hardware is hard but doable. The ecosystem is what kills.

 

[soresu]

I suspect this is somewhat analogous to Huawei being forced to design their own future SOC for their phones post Trump.

They have their own SoC (Kirin), they lack a custom CPU or GPU core that they own.

I would not despair as yet, ARM knows that Huawei is a major customer, they are not sitting idle - even though their latest SoC uses no new cores, I have no doubt that new ones will appear in the next gen.

Huawei skipped a CPU gen with Kirin 970, electing to refine A73 instead, and G72 barely gave them any improvement over G71, so I don't think it's especially surprising that they elected to skip both with K 990.

I still hope to see a next gen SoC with Cortex A78 and Mali G78 next september.

 

[soresu]

Is it worth the (massive) pain on the software ecosystem front?

There is probably more variation of software on iOS than Mac today given the larger market, it would not be so terrible as that.

What they may do is make a Mac/Macbook with either 2 separate SoC's (or a semi custom AMD design) to merge both for the interim.

 

[Thala]

They did but there were far fewer apps available for Mac back in those days, were there not? There's a lot more now, and a lot of big apps like Adobe that would be difficult to bring across.

Not to mention they need to maintain it for x86 and then ARM as well.

That's probably possible, but at that point wouldn't they just look at making their own x86 CPU? Requiring all apps to support ARM would stagnate the app ecosystem, not that it's particularly huge compared to Windows. Developers would need to support both Mac + Mac ARM for years and years.

You have to understand, that it is relatively trivial for developers to target another micro-architecture, as the real work is done by the compiler backend. So no, there is no difficulty compiling existing apps for ARM.
Regarding maintenance of two versions - there is only a single version of source code. The slight effort increase comes from testing both binary variants.

 

[soresu]

You have to understand, that it is relatively trivial for developers to target another micro-architecture, as the real work is done by the compiler backend. So no, there is no difficulty compiling existing apps for ARM.

If done from scratch with the intention of making it suitable for hardware cross platform compatibility, then yes.

However, the emulator PCSX2 has repeatedly asserted that the sheer amount of x86 assembly code written for it would make it extremely difficult to add an ARM back end.

Whether that was always just an excuse for laziness I don't know, but I certainly don't think you have the full grasp of the problem if you think the compiler is the only issue.

I might add that at least one PCSX2 developer is helping to develop DobieStation PS2 emulator, which is aimed at a more cross platform modern Vulkan-ised design - I presume from this that this developer thought it easier to start again from scratch than to retrofit PCSX2.

 

[NTMBK]

What does Apple gain by going to ARM? Longer battery life? More control over the logistics/suppliers? How much of a performance or design advantage could they get with ARM, scaled up to desktop/laptop performance and complexity levels? Is it worth the (massive) pain on the software ecosystem front?

-Only need to maintain ARM toolchains and OS (no more x86 development)- MacOS and iOS have a lot of common components
-No longer indirectly helping their rivals by working with Intel. (Remember Ultrabook, when Intel pushed Windows vendors to copy the Macbook Air?)
-No longer trapped by Intel's inability to deliver. Apple have been stuck with 14nm Skylake derivatives since 2015. In that time Apple's chips have gone from the A9 to the A13, which is a huge jump in performance.

 

[soresu]

Regarding maintenance of two versions - there is only a single version of source code. The slight effort increase comes from testing both binary variants.

Testing on ARM platforms is also tricky because of such a great variance in SoC's and their constituent parts.

GPU's seem to be notoriously bad for ARM platforms, but apparently Samsung M3 had a CPU bug too.

Though obviously on Apple iOS platforms the hardware variance is a lot more regular and known, it should also be noted that anyone developing software from Mac to iOS will at least consider porting to Android if the transition requires significant effort (ie maximise their ROI for porting to ARM).

 

[Thala]

-Only need to maintain ARM toolchains and OS (no more x86 development)- MacOS and iOS have a lot of common components
-No longer indirectly helping their rivals by working with Intel. (Remember Ultrabook, when Intel pushed Windows vendors to copy the Macbook Air?)

In addition they can establish the brand recognition as the highest performing notebooks in their respective category. For example they could easily put 8 A13 cores into a 15W TDP SoC - demolishing anything Intel or AMD has to offer in this power range.

 

[Richie Rich]

What does Apple gain by going to ARM? Longer battery life? More control over the logistics/suppliers? How much of a performance or design advantage could they get with ARM, scaled up to desktop/laptop performance and complexity levels? Is it worth the (massive) pain on the software ecosystem front?

I suspect this is somewhat analogous to Huawei being forced to design their own future SOC for their phones post Trump. The hardware is hard but doable. The ecosystem is what kills.

1. Intel is behind in manufacturing process... Intel's stuck at 14nm and 10nm... Apple's own old A12 was 7nm, new A13 is N7P already
2. Intel is behind in IPC (Apple A13 is most advanced CPU core on market)
3. x86 is power hungry for laptops ... you need bigger and more expensive battery (ARM saves money or expands work-time)
4. x86 power inefficiency limits performance within TDP ... crucial for multi-core workload ... it will be much more important in near future when they get to 3nm and further (local overheating from high power per mm sq).
5. using Intel means no control and no competitive advantage
6. unification of architecture ... only ARM
7. unification of OS ... possible future use one OS core for iOS and MacOS, same drivers, same HW ... just different lightweight GUI

It was kind of inevitable. Apple has a bad experience with PowerPC. They want to leave sinking boat before it cause any damage.

 

[Mopetar]

They did but there were far fewer apps available for Mac back in those days, were there not? There's a lot more now, and a lot of big apps like Adobe that would be difficult to bring across.

I think XCode will already compile applications for both architectures so it's not difficult. The biggest problems would be software that takes advantage of or is designed around architecture specific features or quirks.

The other side of this is that if Apple is building their own ARM SoC they can always add in some hardware accelerators or special instructions of their own that those companies would want to target.

Imo if anything, Apple is more likely to build a x86 CPU then they are to use an ARM CPU as a Mac's processor.

Why would Intel sell them a license and why would Apple even want to pay what something like that would cost? They'd probably need a separate license from AMD as well for the 64-bit instruction set. Then they'd have to design an entirely new CPU that goes into a few millions of Macs each year instead of hundreds of millions of iDevices.

If they do anything I suspect it's just a straight jump to ARM with some of their hardware models and a gradual shift over time. That's the easiest and requires no dependence on other companies, their road maps, etc. It wouldn't be the first time Apple has done something like this so they have a pretty good idea of what the problems will be and how to handle them.

Partnering with someone like AMD to make a hybrid chip is an outside chance at best. There are some reasons for something like that to happen, but there are plenty not to do it as well. But I'd still rate it as far more likely than Apple building their own x86 CPU.

 

[amrnuke]

It's not bad idea. 12xALUs with SMT2 would be smart alternative to current Apple A13 6xALU with no SMT.

I hope you are continuing in the line of sarcasm with this comment...

 

[soresu]

It was kind of inevitable. Apple has a bad experience with PowerPC. They want to leave sinking boat before it cause any damage.

x86 is not sinking by any means now that AMD is finally throwing its weight around, though they've changed places with Intel in terms of stagnation, so until Intel comes back to form the market could tilt towards AMD some, hopefully Lisa doesn't let that go to her head (or the shareholders complacently ask for scaling back R&D so they can enjoy profits).

As for PowerPC, what bad experience do you speak of?

When Apple transitioned to x86, sure PowerPC CPU IPC was not stellar next to x86 - but that hardly warrants a label of bad experience.

A bad experience is more like Microsoft and nVidia with the first Xbox, or Sony with nVidia on PS3, or Nintendo with Switch TX1 security bug.

Basically anyone with nVidia....

 

[Kenmitch]

AM I the only one who doesn't Apple? I could really care less what they do, but at least reading the comments in Apple related threads can be amusing.

 

[scannall]

It was kind of inevitable. Apple has a bad experience with PowerPC. They want to leave sinking boat before it cause any damage.

At the time Apple adopted the PowerPC is was by far faster than anything else, particularly the G5 But then IBM let it stagnate, and refused to work on a low power chip. So it was either make the jump to Intel and x86 or wait until the end of time for IBM to make what they needed. Intel has been stagnant for a while now, which likely has Apple very unhappy. They may well do their own custom CPU, they certainly have the talent. Or they may start introducing AMD parts into some of their lines.

 

[Asterox]

Some random guys youtube channel isn't exactly definitive...

Well It is not Linus Tech Tips, and does not release video every day etc.

What is definitive, it is very intresting rumor or not a crazy rumor from Klingon.

News - ARM Matterhorn

Found this on the ARM TechCon blog: Someone mentioned this to me recently as something related to ARMv9.0-A, possibly the post Hercules big core. Link for the ARM TechCon blog here. A quote under another image: "MatMuil will double CPU GEMM performance" "Matternhorn will introduce what Arm...

forums.anandtech.com

 

[soresu]

Klingon

Klingon??!!

ON THE STARBURD BOW CAPTAIN!!

 

[soresu]

At the time Apple adopted the PowerPC is was by far faster than anything else, particularly the G5 But then IBM let it stagnate, and refused to work on a low power chip. So it was either make the jump to Intel and x86 or wait until the end of time for IBM to make what they needed. Intel has been stagnant for a while now, which likely has Apple very unhappy. They may well do their own custom CPU, they certainly have the talent. Or they may start introducing AMD parts into some of their lines.

Oddly the design firm that Apple acquired along with Keller was involved in a low power POWER design.

 

[scannall]

Oddly the design firm that Apple acquired along with Keller was involved in a low power POWER design.

Yes, PA Semiconductor. The founder of that company is an interesting guy with quite a resume'. I think by that time, Apple was pretty tired of delays from IBM and didn't want to stay tied to a small producer. It would have been interesting to see where that went though.