You're Apple. You just built a 40-core SoC for Mac Pro. Now what?

[senttoschool]

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022.

You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep the Mac Pro relevant. How do you recuperate some of this money spent?

You create "Apple Cloud". No, not iCloud. Apple Cloud. Like AWS. Where anyone can come and rent a 40-core M3 SoC running on macCloudOS. You get into the cloud hosting business. You file this under the "Services" strategy that you keep pushing to make Wall Street happy.

Soon, you'll be releasing 64-core SoCs with 128-core GPUs, then 128-core SoCs with 256-core GPUs, and so on. Somehow, you're actually beating anything AWS, Azure, Google Cloud can offer... without really trying.

Apple Silicon Cloud.

It wouldn't surprise me if Apple is already testing their own SoCs to power their iCloud service, which currently depend on AWS. Apple was reportedly spending $30m/month on AWS in 2019. It might be $100m+ per month by now given how fast services have grown.

 

[ader42]

It could be nice for 3D rendering in the cloud.

 

[jdb8167]

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022.

You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep the Mac Pro relevant. How do you recuperate some of this money spent?

You create "Apple Cloud". No, not iCloud. Apple Cloud. Like AWS. Where anyone can come and rent a 40-core M3 SoC running on macCloudOS. You get into the cloud hosting business. You file this under the "Services" strategy that you keep pushing to make Wall Street happy.

Soon, you'll be releasing 64-core SoCs with 128-core GPUs, then 128-core SoCs with 256-core GPUs, and so on. Somehow, you're actually beating anything AWS, Azure, Google Cloud can offer... without really trying.

Apple Silicon Cloud.

It wouldn't surprise me if Apple is already testing their own SoCs to power their iCloud service, which currently depend on AWS. Apple was reportedly spending $30m/month on AWS in 2019. It might be $100m+ per month by now given how fast services have grown.

I suspect the starting point is Xcode Cloud. It will be interesting to see how it evolves.

 

[bobmans]

It’s not like the “40 core mac pro chip” or whatever demands a lot of r&d specific to that chip. Apple designs it’s chips so that they scale easily. Most of r&d goes into designing the cores and Apple will definitely use the same cores across their entire mac lineup.

 

[Internaut]

If Apple gets far enough ahead of the competition, they may well have 3-4 years. A Mac Cloud? Apple suffers from being a ’shiny new toy first’ kind of a company while Microsoft and Google have made off with much of the services revenue. I’m not an investor, but if I was I think I’d want Apple to be thinking out of the box with the M line.

M1 was easy: get the same damned thing into as many devices as you can. The scale is impressive. When it comes to workstation class processors, it’s not clear that Apple has any prospect of out-scaling Intel. Interesting times; exciting products to come.

 

[Gnattu]

Apple designs it’s chips so that they scale easily.

Not quite. The A14/M1 design is not scalable at all, but I'm looking forward to see Apple's approach to make larger chips.

 

[Gnattu]

I suspect the starting point is Xcode Cloud. It will be interesting to see how it evolves.

That uses Intel Xeon Gold at the moment, but the demand to have an adequate CI/CD for Apple platforms is huge, lots of company bought Mac Minis for this purpose.

However, Apple changed the ToS of macOS after Big Sur, which does not allow short time renting (less than 24 consecutive hours) of the computing resource, and that hurts public CI/CD heavily because the leasing period of CI/CD is typically only a few minutes. Even GitHub, now owned by Microsoft, cannot provide a useful Big Sur CI for now. Such limitation is anti-competitive and make Apple the only company can provide public macOS CI/CD. I'm not a lawyer so I don't know if that's legal, but at least they should push that new renting policy later, at least after the Xcode Cloud is publicly available (not beta). A lot of projects (mostly community driven) cannot afford a dedicated Apple hardware for CI/CD purpose does not have access to any CI/CD now.

 

[jdb8167]

That uses Intel Xeon Gold at the moment, but the demand to have an adequate CI/CD for Apple platforms is huge, lots of company bought Mac Minis for this purpose.

However, Apple changed the ToS of macOS after Big Sur, which does not allow short time renting (less than 24 consecutive hours) of the computing resource, and that hurts public CI/CD heavily because the leasing period of CI/CD is typically only a few minutes. Even GitHub, now owned by Microsoft, cannot provide a useful Big Sur CI for now. Such limitation is anti-competitive and make Apple the only company can provide public macOS CI/CD. I'm not a lawyer so I don't know if that's legal, but at least they should push that new renting policy later, at least after the Xcode Cloud is publicly available (not beta). A lot of projects (mostly community driven) cannot afford a dedicated Apple hardware for CI/CD purpose does not have access to any CI/CD now.

If you can’t test against an Apple Silicon device then you are not really supplying a complete solution. I suspect that at some point Apple wi’ll provide an test solution against actual Apple Silicon for both MacOS and iOS.

 

[senttoschool]

It’s not like the “40 core mac pro chip” or whatever demands a lot of r&d specific to that chip

It does.

When you start making large chips, you have to invest in things like chiplet designs due to lower yields of a monolithic design. You also have to decouple RAM from the main SoC because you can't possibly glue 128GB/256GB or more RAM into an SoC. Big SoCs also require more PCI-E lanes, more memory bandwidth.

The Mac Pro is a very niche market relative to iPhones, iPads, and low-end Macs. Investing a significant chunk of your Apple Silicon R&D into the smallest market doesn't seem to make sense.

 

[Jorbanead]

It does.

When you start making large chips, you have to invest in things like chiplet designs due to lower yields of a monolithic design. You also have to decouple RAM from the main SoC because you can't possibly glue 128GB/256GB or more RAM into an SoC. Big SoCs also require more PCI-E lanes, more memory bandwidth.

The Mac Pro is a very niche market relative to iPhones, iPads, and low-end Macs. Investing a significant chunk of your Apple Silicon R&D into the smallest market doesn't seem to make sense.

I always figured that they would create an M1X and then slap 4 of them together to reach the 40-core machine. So really most of the RnD was used on the M1X which will go in a bunch of macs.

 

[adib]

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022.

You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep the Mac Pro relevant. How do you recuperate some of this money spent?

You create "Apple Cloud". No, not iCloud. Apple Cloud. Like AWS. Where anyone can come and rent a 40-core M3 SoC running on macCloudOS. You get into the cloud hosting business. You file this under the "Services" strategy that you keep pushing to make Wall Street happy.

Soon, you'll be releasing 64-core SoCs with 128-core GPUs, then 128-core SoCs with 256-core GPUs, and so on. Somehow, you're actually beating anything AWS, Azure, Google Cloud can offer... without really trying.

Apple Silicon Cloud.

It wouldn't surprise me if Apple is already testing their own SoCs to power their iCloud service, which currently depend on AWS. Apple was reportedly spending $30m/month on AWS in 2019. It might be $100m+ per month by now given how fast services have grown.

... hence the upcoming Xcode Cloud... Hopefully the software development effort for that would also trickle down to on-premise macOS servers.

 

[sunny5]

Well, they really need to attract some PC friendly software such as 3D stuff cause they only rely on Nvidia and CUDA. But it's up to Apple if they make a powerful GPU or not.

 

[leman]

I somehow doubt that they have an intention of entering the cloud business. Several reasons: it's not a very cost-efficient use for their high-end chips that are going to be supply limited anyway (much more profitable to just sell them in a Mac Pro with huge margin), that would require them to build and maintain a Linux infrastructure, and, they would have to offer disruptive pricing to make a better value proposition. Their chips are very fast, but that is not the main concern in cloud computing. People usually care much more about cost per request rather than performance per dollar.

Apple will offer more cloud services, that's for sure, but I think they will still use the third-party to run them for the overseeable future — it's cheaper and less hassle for them.

 

[thenewperson]

Build 80-core and 160-core versions ¯\_(ツ)_/¯

 

[senttoschool]

I somehow doubt that they have an intention of entering the cloud business. Several reasons: it's not a very cost-efficient use for their high-end chips that are going to be supply limited anyway (much more profitable to just sell them in a Mac Pro with huge margin), that would require them to build and maintain a Linux infrastructure, and, they would have to offer disruptive pricing to make a better value proposition. Their chips are very fast, but that is not the main concern in cloud computing. People usually care much more about cost per request rather than performance per dollar.

Apple will offer more cloud services, that's for sure, but I think they will still use the third-party to run them for the overseeable future — it's cheaper and less hassle for them.

AWS started out as an internal-only service. Same for Google-cloud.

Apple could start by transitioning its iCloud/Maps/Appstore/Apple One services to its own SoCs. Everything Apple will want to do in the future, including self-driving cars, will need the cloud. Might as well build your own internal cloud.

They wouldn't necessarily run Linux. Apple could start by offering macOS emulation on the cloud with a Mac Pro SoC. Next, Apple could offer a UI-less, compute-focused instances of macOS for those who just want to rent Apple Silicon power. macOS can already run any popular server applications such as MySQL, Postgres, Redis, Kafka, Kubernetes, Tensorflow, Pytorch, Nginx.

Another factor in this is cloud gaming. The gaming industry moving to the cloud, especially demanding AAA games. If Apple wants to catch this wave, it too will need to offer cloud rendered games.

 

[leman]

They wouldn't necessarily run Linux. Apple could start by offering macOS emulation on the cloud with a Mac Pro SoC. Next, Apple could offer a UI-less, compute-focused instances of macOS for those who just want to rent Apple Silicon power. macOS can already run any popular server applications such as MySQL, Postgres, Redis, Kafka, Kubernetes, Tensorflow, Pytorch, Nginx.

Trying to make macOS into an OS suitable for cloud services is a far bigger challenge than making macOS gaming-friendly And being able to run a database software and running it in production are two entirely different things. We live in an age of managed services, serverless architecture, on-demand scalability... I just don't see a way how Apple can disrupt that market. Their hardware is great, but they are utterly lacking in software, and they have been dismantling their server efforts up to the point that they are basically non-existent.

Customer-oriented services like Xcode Cloud is a completely different topic, and I suppose a case could be made for having an Apple Compute Cloud. But that last bit is already a bit problematic wrt. software infrastructure. If they only support Metal I don't see any chance for a service like that getting any traction, the user group is just too small.

 

[sunny5]

Another factor in this is cloud gaming. The gaming industry moving to the cloud, especially demanding AAA games. If Apple wants to catch this wave, it too will need to offer cloud rendered games.

Mac itself is terrible for gaming and all gaming platform is based on Windows. Who even use macOS? You will need a powerful GPU for cloud gaming but do Apple has it? Nope.

The OS market share of macOS is less than 16% and I dont think it will be competitive.

 

[senttoschool]

Mac itself is terrible for gaming and all gaming platform is based on Windows. Who even use macOS? You will need a powerful GPU for cloud gaming but do Apple has it? Nope.

The OS market share of macOS is less than 16% and I dont think it will be competitive.

In 3 years, 50% of all computers capable of playing AAA games will be Macs

Edit: December 13th, 2021: Apple will sell about 27 million Macs in 2021. Assume that 80% are M1/M1Pro/M1Max Macs, then Apple will have sold 21 million Macs capable of playing AAA games. According to IDC, the number of gaming computers (desktops and laptops) sold in 2021 is projected to be 47...

forums.macrumors.com

 

[sunny5]

In 3 years, 50% of all computers capable of playing AAA games will be Macs

Edit: December 13th, 2021: Apple will sell about 27 million Macs in 2021. Assume that 80% are M1/M1Pro/M1Max Macs, then Apple will have sold 21 million Macs capable of playing AAA games. According to IDC, the number of gaming computers (desktops and laptops) sold in 2021 is projected to be 47...

forums.macrumors.com

The problem with your statement is that you only mentioned a hardware capability, not the platform itself. That's why your post is so laughable.

- OS market share is extremely low for macOS.
- Therefore, way less gamers on macOS than Windows.
- Mac itself is expensive than PC for its performance.
- macOS is not friendly for gaming.
- Apple lacks gaming related technology.
- Most PC/Console games are Windows based.
- The market for macOS is really small and not profitable.
- Developers from PC/Console are not interested in ARM.
- And more...

Clearly, you know nothing about the gaming market after all. Having a capable hardware does not mean Apple can increase their gaming market and they are focusing only on mobile.

 

[senttoschool]

- OS market share is extremely low for macOS.

Does not matter. The post looks at the number of computers capable of playing AAA games.

- Therefore, way less gamers on macOS than Windows.

Again, pay attention to the actual post.

- Mac itself is expensive than PC for its performance.

False. You can buy a $300 Costco Windows laptop but it can't play any games. Any Apple Silicon, including a $600 Mac Mini can.

- Apple lacks gaming related technology.

Apple is actually the largest gaming company in the world by profit. By definition, Metal is the most popular developer API. What "gaming related technology" are you referring to?

- Most PC/Console games are Windows based.

This is incorrect. But I think what you're trying to say is that Windows has more AAA games. Regardless, my post looks into the future.

- The market for macOS is really small and not profitable.

Define small. Define "not profitable".

- Developers from PC/Console are not interested in ARM.

Unreal Engine/Unity Engine supports ARM out of the box. There are more ARM-based games than there are x86/x64 games. Developers go where there is a big platform, regardless of the CPU architecture.

- And more...

Let's hear them

Clearly, you know nothing about the gaming market after all.

What's your qualification in gaming?

Having a capable hardware does not mean Apple can increase their gaming market and they are focusing only on mobile.

I suggest you read this post:

In 3 years, 50% of all computers capable of playing AAA games will be Macs

Edit: December 13th, 2021: Apple will sell about 27 million Macs in 2021. Assume that 80% are M1/M1Pro/M1Max Macs, then Apple will have sold 21 million Macs capable of playing AAA games. According to IDC, the number of gaming computers (desktops and laptops) sold in 2021 is projected to be 47...

forums.macrumors.com

Clearly, looking past the present and projecting into the future isn't your strong suit.

 

[sunny5]

Does not matter. The post looks at the number of computers capable of playing AAA games.

It matters lol. How many AAA games are there for Mac? What about Apple Silicon? Being capable doesn't prove anything at all. Also, most recent AAA games aren't even supported on Mac.

Again, pay attention to the actual post.

That's just your opinion, nothing more. Why do I even bother to read your THEORY? Especially since you are not providing any links or information to support your claim? That's not how people to argue.

False. You can buy a $300 Costco Windows laptop but it can't play any games. Any Apple Silicon, including a $600 Mac Mini can.

And that's only one example which you failed to prove your point. Also, Mac does not support Diret X and ray tracing.

Apple is actually the largest gaming company in the world by profit. By definition, Metal is the most popular developer API. What "gaming related technology" are you referring to?

In MOBILE, not PC/Console. They are def a different market and yet you are considering them as a same market. Direct Storage, DirectX, Ray tracing, DLSS, and more.

This is incorrect. But I think what you're trying to say is that Windows has more AAA games. Regardless, my post looks into the future.

Aren't we talking about PC/Console games? You still didn't provide any proofs. Your post is just a theory without proves.

Define small. Define "not profitable".

macOS's market share is less than 16% and how many gamers playing with Mac? You can imagine easily.

Unreal Engine/Unity Engine supports ARM out of the box. There are more ARM-based games than there are x86/x64 games. Developers go where there is a big platform, regardless of the CPU architecture.

So where are PC/Console games on ARM? I said mobile and PC/Console have different market and yet you are making them as a same market.

What's your qualification in gaming?

Steam Hardware & Software Survey

Almost 90% is Windows from Steam. What do you expect? Do you really think macOS is dominant in PC/Console gaming market? Def, not.

Clearly, you know nothing about the gaming market after all.

 

[AltecX]

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022.

You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep the Mac Pro relevant. How do you recuperate some of this money spent?

You create "Apple Cloud". No, not iCloud. Apple Cloud. Like AWS. Where anyone can come and rent a 40-core M3 SoC running on macCloudOS. You get into the cloud hosting business. You file this under the "Services" strategy that you keep pushing to make Wall Street happy.

Soon, you'll be releasing 64-core SoCs with 128-core GPUs, then 128-core SoCs with 256-core GPUs, and so on. Somehow, you're actually beating anything AWS, Azure, Google Cloud can offer... without really trying.

Apple Silicon Cloud.

It wouldn't surprise me if Apple is already testing their own SoCs to power their iCloud service, which currently depend on AWS. Apple was reportedly spending $30m/month on AWS in 2019. It might be $100m+ per month by now given how fast services have grown.

They will do what they always do.... Not upgrade the Mac Pro until its old and a joke compared to PC hardware then about 2-3yrs after that release a new Mac Pro to let sit and rot with barely and updates for another long period of time.

This is what drove me nuts about working in a video studio. We'd be buying Mac Pros that were old and out of date but Apple was still selling as New when for a fraction of the price I could get a nice PC desktop that would destroy it in performance, but it didn't run Final Cut. Finally, even the video guys were so annoyed by this that they started to learn premier just so that they didn't have to be locked into Apple and could occasionally use a PC when they didn't want to drop 5k on a 'new' Mac that came out like 3yrs prior. They would spec out a new PC that was 1/2 the price and then replace it with only minor updates.

 

[robco74]

They just released new GPUs for the Mac Pro. Not sure how that is letting it become an old joke. The trashcan was the exception, not the rule. It's generally Intel that has let their processor line languish, especially Xeon, with few updates that make it worth upgrading. The only reason the 11th gen is worthwhile is that Intel was finally able to hit their new process.

As for gaming, that again falls to Intel. Laptops make up the lion's share of Mac sales. Until now, most of those were smaller laptops with Intel's lackluster iGPUs that were incapable of gaming. Now we have M1 as a baseline, which can manage decent framerates at medium settings and performs better than iGPUs from Intel or AMD. If Apple's GPU performance scales, and there's no reason to think that it won't, it bodes well for higher end products. That means that more Macs are more capable of gaming than ever before.

But there is also the problem of middleware and developer frameworks. From what I understand, this will be an issue for WoA as well. Not an insurmountable problem, but it will take some time. Back in the day, high end software had to be portable as higher end workstations ran different ISAs and OSes. Since then, portability has gone out the window as most people standardized on Wintel. Since that is now changing, it will take some time to adapt.

I don't think the Mac will become a AAA gaming powerhouse overnight, but rather a capable platform that, if sales continue to improve, will be worth it for more game developers to consider.

 

[Jorbanead]

They will do what they always do.... Not upgrade the Mac Pro until its old and a joke compared to PC hardware then about 2-3yrs after that release a new Mac Pro to let sit and rot with barely and updates for another long period of time.

This is what drove me nuts about working in a video studio. We'd be buying Mac Pros that were old and out of date but Apple was still selling as New when for a fraction of the price I could get a nice PC desktop that would destroy it in performance, but it didn't run Final Cut. Finally, even the video guys were so annoyed by this that they started to learn premier just so that they didn't have to be locked into Apple and could occasionally use a PC when they didn't want to drop 5k on a 'new' Mac that came out like 3yrs prior. They would spec out a new PC that was 1/2 the price and then replace it with only minor updates.

Was this pre-2013 or post-2013?

Post-2013 was a **** show and Apple has admitted fault in their design/plan. Before 2013 they updated on a 1-2 year cycle which honestly seemed reasonable. Issue after that was the trash can design, but the new design isn’t constrained. Apple knows post-2013 trash can caused a lot of controversy and issue for professionals so they’re not about to make that same mistake again.

 

[GSWForever8]

Mac itself is terrible for gaming and all gaming platform is based on Windows. Who even use macOS? You will need a powerful GPU for cloud gaming but do Apple has it? Nope.

The OS market share of macOS is less than 16% and I dont think it will be competitive.

Macs aren’t that bad at gaming, it’s just the developer that won’t develop for macs.