No “Extreme” chip coming to Mac Pro?

[Colstan]

I think one full sized though? Or am I misremembering?

Gurman originally said that there will be a smaller and larger Mac Pro. The small version turned out to be the Mac Studio. Now he just talks about the big boy.

I think it was more well the rumor is the Extreme is dead, maybe it’ll have a dGPU as unlikely as it seems - that would really push the need for internal PCIe, no that’s dead too, maybe it’s expandable RAM, no that’s dead too.

There are folks, mainly over at the other place, who insist that the Mac Pro needs third-party graphics cards, because otherwise nobody would buy them. At this point, all of the speculation is turning my brain to mush, and I'm not even in the market for one.

personally I think the notion that Apple will never introduce an extreme is bunk. They will. But it will be unfortunate if it doesn’t come out with the Mac Pro.

My guess is something went wrong with the M2 version, so they are delaying it until M3, unless the bean counters simply told them it wasn't worth bothering with. Tim Cook is pragmatic, which doesn't always align with the nice things that we want.

I just hope Apple is talking to real users; rather than internal EEs/CSEEs who indeed may be very smart and capable, may not understand what real users would really like; and thus make poor strategic decisions.

Apple doubtlessly has far superior marketing data than us mere mortals. Whatever they decide, it will be carefully crafted, both in presentation and specifications, for maximum profit.

There are doubtless engineers at Apple who really like using mac pros, which is probably part of what keeps them coming. As long as Apple doesn’t lose money on them, they’re likely safe.

Excellent point.

But if they are going to get into the habit of making CPUs/SoCs just for Mac Pro, there’s a good chance they could end up losing money on them.

Which is the $5,999 question. I would like to know how many they actually sell, because then we might get a glimpse about how much it matters to the fruit company.

 

[dada_dave]

Gurman originally said that there will be a smaller and larger Mac Pro. The small version turned out to be the Mac Studio. Now he just talks about the big boy.

Sorry I meant only one full sized PCIe slot. Originally I think the rumor was that even the bigger Pro was still going to have a smaller chassis than the current one.

There are folks, mainly over at the other place, who insist that the Mac Pro needs third-party graphics cards, because otherwise nobody would buy them. At this point, all of the speculation is turning my brain to mush, and I'm not even in the market for one.

My guess is something went wrong with the M2 version, so they are delaying it until M3, unless the bean counters simply told them it wasn't worth bothering with. Tim Cook is pragmatic, which doesn't always align with the nice things that we want.

Oh I agree about the rumors - It’s fun to theorize but beyond that people get too bent out of shape.

As for the extreme’s existence, I dunno as I don’t put much stock in these rumors except those that make sense with what we’ve seen so far: a dGPU and expandable RAM would difficult under the current product model ethos. So it makes sense that they wouldn’t be offered. I won’t say it’s impossible that Apple failed to create an M2 extreme but I’ll wait until the products are released before believing it. If that is the case I can see them delaying the entire Mac Pro line until the M3. Hell we don’t even know if that wasn’t the plan all along and there never was an M2 Extreme in the internal roadmap!

Apple doubtlessly has far superior marketing data than us mere mortals. Whatever they decide, it will be carefully crafted, both in presentation and specifications, for maximum profit.

And occasionally they get it wrong - even in the fabled Jobs era there were duds and missed opportunities.

 

[Citysnaps]

Apple doubtlessly has far superior marketing data than us mere mortals. Whatever they decide, it will be carefully crafted, both in presentation and specifications, for maximum profit.

Imo... from working at a few tech companies, I think it could be a mixed bag.

Worst outcomes, imo, while working for a large semiconductor company, formal marketing engaged customers which resulted and drove product development. Best outcome was where systems engineers well-versed in communications theory and chip design engaged customers who were also engineers and managers - both sides spoke the same language, brainstorm, could set specifications and requirements, deal with in-depth technical questions on the spot, etc - everything was smooth.

Working at a tiny company chip company before it was acquired by the above semiconductor company, there were always great outcomes, because everyone was a systems and hardware engineer, designed chips, and spoke the same language as engineers at companies we engaged. We had no real marketing department. Customer engineers loved talking to us because we spoke the same language, could brainstorm technical solutions on the spot, set requirements and timelines, etc. Always smooth. Customers appreciated that.

Before that... the above employees came from a medium sized defense aerospace company that developed complex systems relating to signals acquisition and analysis. Even special chips. Systems engineers visited customers, after that may have done some initial IR&D for feasibility on a potential project, and then wrote a proposal for the customer. Very little formal marketing involved, other than perhaps a high-level $$$ sanity check. Customers liked that because they and company system engineers spoke the same language.

Sorry for the long story... but that's what drove me wondering how product development works at Apple.

Seems there could be a wide range of possibilities leading to what kind of new MacPro is developed. Which could ultimately drive success in the market. IMO, it's all about having technical people engage a wide range of real life experienced users and listening to their needs and requirements. If I were Apple I would certainly want to talk to Neil Parfitt and others like him - both in sound/music fields as well as other disciplines.

 

[Colstan]

Sorry I meant only one full sized PCIe slot. Originally I think the rumor was that even the bigger Pro was still going to have a smaller chassis than the current one.

Same case, same funky holes.

Oh I agree about the rumors - It’s fun to theorize but beyond that people get too bent out of shape.

After the official announcement, I'm getting the popcorn out, and heading over the the MR forums. I can't wait.

If that is the case I can see them delaying the entire Mac Pro line until the M3. Hell we don’t even know if that wasn’t the plan all along and there never was an M2 Extreme in the internal roadmap!

Assuming that the "Extreme" ever existed, I don't think they'll delay the Apple Silicon model until M3. I think it will be, at best, first released with an M2 Ultra. Then, with the M3, we may also see an "Extreme" variant. Apple isn't waiting another year or more to update the Mac Pro. They want x86 dead, tossed in a shallow grave, covered with quicklime, and buried under an obelisk. The sooner the Apple Silicon Mac Pro is released, the sooner they can accomplish that. The product isn't important enough to wait it out for another M-series generation.

 

[dada_dave]

Same case, same funky holes.

After the official announcement, I'm getting the popcorn out, and heading over the the MR forums. I can't wait.

Assuming that the "Extreme" ever existed, I don't think they'll delay the Apple Silicon model until M3. I think it will be, at best, first released with an M2 Ultra. Then, with the M3, we may also see an "Extreme" variant. Apple isn't waiting another year or more to update the Mac Pro. They want x86 dead, tossed in a shallow grave, covered with quicklime, and buried under an obelisk. The sooner the Apple Silicon Mac Pro is released, the sooner they can accomplish that. The product isn't important enough to wait it out for another M-series generation.

Not unless the M3 extreme comes out as the first in its generation - it wouldn’t fit Apple’s MO I grant you but it wouldn’t be the first time a company comes out with the largest die/product first - it’s often the case in the dGPU market.

 

[Cmaier]

Same case, same funky holes.

After the official announcement, I'm getting the popcorn out, and heading over the the MR forums. I can't wait.

Assuming that the "Extreme" ever existed, I don't think they'll delay the Apple Silicon model until M3. I think it will be, at best, first released with an M2 Ultra. Then, with the M3, we may also see an "Extreme" variant. Apple isn't waiting another year or more to update the Mac Pro. They want x86 dead, tossed in a shallow grave, covered with quicklime, and buried under an obelisk. The sooner the Apple Silicon Mac Pro is released, the sooner they can accomplish that. The product isn't important enough to wait it out for another M-series generation.

The fact that there don’t seem to be any pictures around of an unpackaged M2 Max die gives me hope that they’re hiding crossbars on two edges (which would support Extreme). The only die photos I’ve seen are from Keynote slides, where things could be cropped out. In fact, I don’t recall seeing a die photo with even one crossbar, though I may have missed it, and we know there has to be at least one.

 

[Citysnaps]

After the official announcement, I'm getting the popcorn out, and heading over the the MR forums. I can't wait.

Same here.

Go for the pontificating: "What real PROs really need is blah blah bah."

Stay for the race-to-the-bottom inane whine-fests because Apple is stupid.

 

[theorist9]

Imo... from working at a few tech companies, I think it could be a mixed bag.

Worst outcomes, imo, while working for a large semiconductor company, formal marketing engaged customers which resulted and drove product development. Best outcome was where systems engineers well-versed in communications theory and chip design engaged customers who were also engineers and managers - both sides spoke the same language, brainstorm, could set specifications and requirements, deal with in-depth technical questions on the spot, etc - everything was smooth.

Working at a tiny company chip company before it was acquired by the above semiconductor company, there were always great outcomes, because everyone was a systems and hardware engineer, designed chips, and spoke the same language as engineers at companies we engaged. We had no real marketing department. Customer engineers loved talking to us because we spoke the same language, could brainstorm technical solutions on the spot, set requirements and timelines, etc. Always smooth. Customers appreciated that.

Before that... the above employees came from a medium sized defense aerospace company that developed complex systems relating to signals acquisition and analysis. Even special chips. Systems engineers visited customers, after that may have done some initial IR&D for feasibility on a potential project, and then wrote a proposal for the customer. Very little formal marketing involved, other than perhaps a high-level $$$ sanity check. Customers liked that because they and company system engineers spoke the same language.

Sorry for the long story... but that's what drove me wondering how product development works at Apple.

Seems there could be a wide range of possibilities leading to what kind of new MacPro is developed. Which could ultimately drive success in the market. IMO, it's all about having technical people engage a wide range of real life experienced users and listening to their needs and requirements. If I were Apple I would certainly want to talk to Neil Parfitt and others like him - both in sound/music fields as well as other disciplines.

Apple brought in a "Pro Workflow Team" comprised of industry professionals to help them design the 2019 Mac Pro (and avoid the mistakes with the 2013 model):

Apple modular Mac Pro launch coming in 2019, new engineering group formed to guarantee future of hardware | AppleInsider

A redesigned modular Mac Pro — teased in April 2017 for professionals that want to upgrade faster — won't ship until 2019, Apple declared on Thursday.

appleinsider.com

Mac Pro's lessons learned will trickle down to all 'Pro' products, says project lead | AppleInsider

The product manager for this fall's 2019 Mac Pro, Doug Brooks, promised in a podcast interview this week that the impact of his Pro Workflow Team will be felt beyond this year's WWDC centerpiece.

appleinsider.com

According to Doug Brooks, the product manager for the 2019 Mac Pro, "functionally it's split into "two kinds of people," ...the first are "artists and creatives who came out of the industry, in the video, audio, and 3D space," while the second are "system architecture people" and "deep computer scientists" that partner with the first group.

Unfortunately, I've not seen any mention of including natural scientists (physicists, chemists, biologists, geologists, astronomers, meterologists, etc.).

This group was also involved in the redesign of the last Intel MBPs (you'll recall these departed from the designs of Jony Ive era, as they were thicker to give them better thermals), as well as the AS MBP's and Mac Studio:

Mac Studio designed with customer feedback & Pro Workflow team feedback in mind | AppleInsider

The Mac Studio and Studio Display were built with pros in mind, say Apple executives, because modularity, connectivity, and performance are what matter most.

appleinsider.com

I'd expect they are also giving Apple detailed input for the upcoming AS MP. Supposedly they played a role in driving the modularity you see in the 2019 MP. Of course, design constraints may prevent them from getting that with the AS MP.

 

[Colstan]

Same here.

Go for the pontificating: "What real PROs really need is blah blah bah."

It's going to be fantastic.

Stay for the race-to-the-bottom inane whine-fests because Apple is stupid.

Don't forget "Apple is doomed" and "Tim Cook needs to be fired".

 

[Citysnaps]

Apple brought in a "Pro Workflow Team" comprised of industry professionals to help them design the 2019 Mac Pro (and avoid the mistakes with the 2013 model):

Apple modular Mac Pro launch coming in 2019, new engineering group formed to guarantee future of hardware | AppleInsider

A redesigned modular Mac Pro — teased in April 2017 for professionals that want to upgrade faster — won't ship until 2019, Apple declared on Thursday.

appleinsider.com

Mac Pro's lessons learned will trickle down to all 'Pro' products, says project lead | AppleInsider

The product manager for this fall's 2019 Mac Pro, Doug Brooks, promised in a podcast interview this week that the impact of his Pro Workflow Team will be felt beyond this year's WWDC centerpiece.

appleinsider.com

According to Doug Brooks, the product manager for the 2019 Mac Pro, "functionally it's split into "two kinds of people," ...the first are "artists and creatives who came out of the industry, in the video, audio, and 3D space," while the second are "system architecture people" and "deep computer scientists" that partner with the first group.

Unfortunately, I've not seen any mention of including natural scientists (physicists, chemists, biologists, geologists, astronomers, meterologists, etc.).

This group was also involved in the redesign of the last Intel MBPs (you'll recall these departed from the designs of Jony Ive era, as they were thicker to give them better thermals), as well as the AS MBP's and Mac Studio:

Mac Studio designed with customer feedback & Pro Workflow team feedback in mind | AppleInsider

The Mac Studio and Studio Display were built with pros in mind, say Apple executives, because modularity, connectivity, and performance are what matter most.

appleinsider.com

I'd expect they are also giving Apple detailed input for the upcoming AS MP. Supposedly they played a role in driving the modularity you see in the 2019 MP. Of course, design constraints may prevent them from getting that with the AS MP.

Great find! And good to hear. No doubt something similar is/was organized for the new MacPro.

 

[Citysnaps]

It's going to be fantastic.

Don't forget "Apple is doomed" and "Tim Cook needs to be fired".

That's always a good go-to crowd-pleaser!

 

[Citysnaps]

Unfortunately, I've not seen any mention of including natural scientists (physicists, chemists, biologists, geologists, astronomers, meterologists, etc.).

That would have been great. Medical research too.

 

[Andropov]

Using fell magic, dark arts, and a scrying pool, I've been able to divine the answer concerning the expandability of the Apple Silicon Mac Pro, if Apple decides to no longer support third-party graphics cards, which currently seems to be the case. I have finally found the answer.

Exhibit "A" are the Sonnet Echo I and Echo III expansion modules. If you go down to the section that says "Versatile PCIe Card Expansion" it will have a link to a PDF for supported PCIe cards. So that you don't have to wade through that page, here is the direct link for the relevant PDF. Take a gander at "M1/M2 Mac Compatible Driver". There are currently 60+ PCIe cards that are supported by Apple Silicon. None of those are GPUs.

So, even if Apple doesn't provide graphics card support for the next Mac Pro, there is certainly plenty reason to have PCIe slots.

IMHO, Apple is less focused in what the market is currently doing and more in what the market should be ideally doing. Are there hundreds of useful PCIe cards out there that all sorts of professionals are using? Absolutely. Is PCIe the right standard for those cards or just a vestige from the past? I'm less sure about that. Apple concluded otherwise at least once in the past (with the 2013 Mac Pro). And Apple has been trying to kill the 'Pro' tower since... forever. The G4 Cube was certainly not the right time, the 2013 Mac Pro was killed by the failure of OpenCL and multi-GPU systems... Nowadays Apple seems to be more conservative with its Pro systems, but still, I think Apple will only spare the PCIe ports if they have benefits over other technologies (like Thunderbolt) other than third party support.

The fact that there don’t seem to be any pictures around of an unpackaged M2 Max die gives me hope that they’re hiding crossbars on two edges (which would support Extreme). The only die photos I’ve seen are from Keynote slides, where things could be cropped out. In fact, I don’t recall seeing a die photo with even one crossbar, though I may have missed it, and we know there has to be at least one.

They crop crossbars out of the die shots, yeah. The M1 Max in its keynote's die shots didn't show a crossbar either.

 

[Buntschwalbe]

As i don't have the technical insight to join your discussions about how a Mac Pro would look like on the inside, i'll try to approach the topic from a different angle.

My assumption is, that apple is still fully focused on its transition to Apple silicon. Yes, Delayed , but still they will try to move away from intel based systems as fast as possible. Only the Mac Pro is missing out that transformation yet. So i predict, that apple will release a M2 update of the Mac Studio and the Imac 24" (maybe even 27"?) this spring and will announce a Mac Pro (M3 ultra/extreme) at the WWDC to be released in fall/winter 2023. This will increase the time passing between the release of M2 Products and the next iteration of its Mx chips. There needs to be some gap between M2 and M3, otherwise freshly updated devices will look obsolete a few months after their release.

I do miss a lot the software part, on the discussion about apples transition to Arm. I think their two Year window was also supposed to be the time Periode in witch the ditch intel support on MacOs. To save money and simplify their development they'll make MacOs14 Apple silicon only.

I know there are a lot of guesses written above, and clearly i cannot predict the future of Apples developments. I hope, you forgive me my boldness. My only Point is that we can approach this topic from a different perspective. Looking at apple as a company, acting as rational player to maximise profits and minimise costs of production and R&D; trying to have a steady and well timed output of new software and hardware releases.

 

[Citysnaps]

I haven't looked at FPGAs in a long time (I see Altera is now under Intel, and Xilinx is part of AMD), but perhaps a PCIe card with a couple or few high-end Xilinx FPGAs used as accelerators would be interesting. An initial thought, haven't played around with numbers or thought about applications.

 

[exoticspice1]

@exoticspice1 Btw I was slightly wrong here. This is an Apple style ultra fusion chiplet design. I got confused by the reuse of the nvlink name but this has a similar (slightly higher) bandwidth as ultra fusion and seems to be similar upon closer reading. So this is the overall style in which Apple probably would have/will make an “Mx extreme” chip but bigger as it is meant for data center/clusters. I doubt Apple will ever create something quite that big as I they tend to be consumer focused.

Mm interesting. Thanks for the additional info.

 

[quarkysg]

The fact that there don’t seem to be any pictures around of an unpackaged M2 Max die gives me hope that they’re hiding crossbars on two edges (which would support Extreme). The only die photos I’ve seen are from Keynote slides, where things could be cropped out. In fact, I don’t recall seeing a die photo with even one crossbar, though I may have missed it, and we know there has to be at least one.

What is the possibility of placing the SoC on both sides of the logic boards? Would leave more space for memory modules on each sides, and only need one side of the SoC to connect to the UltraFusion derivative.

Probably a nightmare to manufacture if its even possible with so many signals to go thru the board.

 

[Cmaier]

What is the possibility of placing the SoC on both sides of the logic boards? Would leave more space for memory modules on each sides, and only need one side of the SoC to connect to the UltraFusion derivative.

Probably a nightmare to manufacture if its even possible with so many signals to go thru the board.

Very unlikely.

 

[B01L]

Assuming that the "Extreme" ever existed, I don't think they'll delay the Apple Silicon model until M3. I think it will be, at best, first released with an M2 Ultra. Then, with the M3, we may also see an "Extreme" variant. Apple isn't waiting another year or more to update the Mac Pro. They want x86 dead, tossed in a shallow grave, covered with quicklime, and buried under an obelisk. The sooner the Apple Silicon Mac Pro is released, the sooner they can accomplish that. The product isn't important enough to wait it out for another M-series generation.

Not unless the M3 extreme comes out as the first in its generation - it wouldn’t fit Apple’s MO I grant you but it wouldn’t be the first time a company comes out with the largest die/product first - it’s often the case in the dGPU market.

The fact that there don’t seem to be any pictures around of an unpackaged M2 Max die gives me hope that they’re hiding crossbars on two edges (which would support Extreme). The only die photos I’ve seen are from Keynote slides, where things could be cropped out. In fact, I don’t recall seeing a die photo with even one crossbar, though I may have missed it, and we know there has to be at least one.

They crop crossbars out of the die shots, yeah. The M1 Max in its keynote's die shots didn't show a crossbar either.

...will announce a Mac Pro (M3 ultra/extreme) at the WWDC to be released in fall/winter 2023. This will increase the time passing between the release of M2 Products and the next iteration of its Mx chips. There needs to be some gap between M2 and M3, otherwise freshly updated devices will look obsolete a few months after their release.

Maybe we have not seen any crossbars on the M2 Max yet because there aren't any, the M2 Max is strictly for the MacBook Pros...

Maybe the M3 Max has the crossbars, maybe two crossbars for an Extreme (4-way) configuration...

Maybe these variant SoCs have higher clocks/voltage available for higher performance and Apple is keeping those dies to the Mac Studio & Mac Pro...

Maybe Apple introduces the high-end of the M3 SoC family with the Mac Pro in WWDC 2023 , "Shipping next week..."; that's how they get all the pre-orders, the phrase engages people's FOMO mode...

 

[leman]

Given the 64-bit address space, it would be an interesting approach. Some Australian university tried it a couple decades ago – a unified mapping that, presumably, only made select regions visible to a given process. They said it offered accelerated task launch, and obviously would simplify inter-process communication. I would imagine it could also provide for fixed address space for code (this particular app always loads at 0xFFB7400851331080). Their idea was to have files loading into the same spot every time, so that literal pointers could be used in the file, about which I would be skeptical, but the basic idea seems sound.

Nvidia has been offering this for years with CUDA. What they do is reserve a huge chunk of process virtual memory and then use the same VM memory ranges on the GPU side. With Apple it would be even simpler — you'd just need to map the same memory pages on both the CPU and the GPU. We already know that Apple actually does it in their kernel interface (based on the Asahi GPU work).

But of course there are a lot of low-level details that are messy. Like flushing GPU caches or memory page structure/format. For example, it might make sense to use larger page sizes for the GPU (16KB is probably too granular). Either way, something that needs to be solved.