MP All Models Is the Mac Pro still relevant in 2025

[Happy_John]

y

I own two 2019 Mac Pros, and very nice to work on.

Yes, they are joy to open up and add / exchange parts. Purely as an enclosure / case, it’s probably the bicepsy I’ve ever for my grubby hands on, and I include any PC enclosure I’ve worked on. It’s incredibly well designed for the user to add components to.

The 5,1 was also lovely for hands-on work, but a earlier time, not not quite as fancy. But far better designed than the PC cases of the time.

IF I remember rightly, the only reason iFixit didn’t give it 10/10 was the proprietary SSD and the T2 chip needing the DFU restre process.

 

[mattspace]

So you're saying Apple should have shipped t-ships with every nMP!

Maybe if the engineering, design, and product marketing teams had been required to wear them every day until they fixed their mistake, there might have been a 2014, or 2015 Mac Pro.

Today, there is nothing that an Apple MacBook Pro can’t handle. Trust me, I used to be a desktop avid desktop user. This new generation of Apple Silicone MacBook Pro will slice through anything you throw at it no matter your discipline. Plus they’re portable.

Driving 12, or 16 4k displays, at full frame rate, full 10 bit colour.

I can do it for 12 right now, and still have ports left over. Any of the folks here with dual W6800 Duos can do it for 16.

Your macbook can do at best 4, the most expensive Studio or M2 Mac Pro maxes out at 6, or 8.

 

[iPadified]

Because the "triangular thermal heat dissipation unit" is the only reason that the trashcan was, well, trashcan-shaped!

What you're asking for is just a circular (for no good reason) Mac Studio.

Exactly! Reason: Beautiful and unobtrusive. Mac Studio is not, there is no imagination there and screams computer. Form matters or else we would still have only towers (for no good reason).

Edit: Reason number 2: it was build like a chimney with likely better passive cooling than the Mac Studio.

 

[iPadified]

Driving 12, or 16 4k displays, at full frame rate, full 10 bit colour.

Probably a 1% of the 1% that need the current Mac Pro needs this. Seems like a very special setup as it is a wall of computer screens. If it is corporate, large government or industries/powerplants, they likely have Win boxes anyway.

 

[mattspace]

Probably a 1% of the 1% that need the current Mac Pro needs this. Seems like a very special setup as it is a wall of computer screens. If it is corporate, large government or industries/powerplants, they likely have Win boxes anyway.

Ahem:

Today, there is nothing that an Apple MacBook Pro can’t handle.

Handle. My. Workflow.

 

[frou]

Is the triangular core and dual-GPU nature of the 6,1 really that big a deal? If there were only 2 hot components, couldn't a heatpipe just be used make contact with the third face?

 

[iPadified]

Driving 12, or 16 4k displays, at full frame rate, full 10 bit colour.

Probably a 1% of the 1% that need the current Mac Pro needs this. Seems like a very special setup as it is a wall of computer screens. If it is corporate, large government or industries/powerplants, they likely have Win boxes anyway.

Ahem:



Handle. My. Workflow.

The WF of one person is completely irrelevant for a business decision.

 

[theluggage]

Ahem:

So, one person in the entire debate makes an overblown comment....

Handle. My. Workflow.

You mean design a whole new Apple Silicon die which threw away some of the biggest advantages of Apple Silicon to support multiple 16-lane PCIe GPUs Just. For. Me. Or keep supporting Mac OS on x86 Just For Me.

It would be very nice of Apple to do that out of appreciation for the customer loyalty of that fraction of a percent - but it ain't gonna happen.

A PC manufacturer might be able to produce the "perfect" video wall machine, because they are free to mix and match from a huge range of commodity parts.

The problem with the Mac Pro is that even the 2019 "peak Mac Pro" only covered a fraction of the PC workstation/server workspace. There were "Scalable Xeon" solutions with more CPU cores, arguably better-than-Xeon processors from AMD, specialist multi-GPU systems that could take even more GPUs, proper rack-mount systems (with redundant PSUs, lights-out management, hot-swappable storage etc...)... Then, at the lower end, if you just needed some PCIe slots or had to have ECC RAM there were far cheaper PC systems that offered that. It was a "one size fits all" product - albeit an extra-stretchy one - in a price range where requirements were increasingly specialised. Apart from a short hineymoon period where it was one of the first to use the new Xeon-W with extra PCIe and RAM capacity, its only selling points were MacOS (in a market segment where Linux was growing and an increasing number of important software tools were multi-platform), a grotesquely over-engineered steampunk case (if that floats your boat) and a slightly neater system for routing power and thunderbolt to GPU cards.

It was telling, at the time, that all of Apple's headline benchmarks were comparing the Mac Pro to Trashcans and iMac Pros rather than PCs. The target market was clearly a short-term plan to retain customers who were locked-in to a MacOS work flow. That's not a recipe for growth, or a good reason to invest in developing new silicon.

It's possible that we'll see a new Apple processor - targetted on AI training/development, but I suspect a key factor in that will be getting enough LPDDR RAM piggybacked on an Apple Silicon SoC so the GPU, CPU and neural engine will be able to share a shedload of unified RAM. But even if that ends up in a box with "Mac Pro" on the front I don't think its gonna be a solution to your x86+PCIe needs.

 

[m1maverick]

Driving 12, or 16 4k displays, at full frame rate, full 10 bit colour.

I can do it for 12 right now, and still have ports left over. Any of the folks here with dual W6800 Duos can do it for 16.

Your macbook can do at best 4, the most expensive Studio or M2 Mac Pro maxes out at 6, or 8.

I suspect this is the reason the OP started this thread. The current Mac Pro, despite having PCIe slots, cannot do what the previous Mac Pro can do.

 

[mattspace]

The WF of one person is completely irrelevant for a business decision.

It's less that I expect Apple to do this, than I expect people who don't own Mac Pros, and don't use Mac Pros to stop with the "you can do anything on a Studio / Macbook" proclamations, as if they have any clue what they're talking about.

 

[mattspace]

You mean design a whole new Apple Silicon die which threw away some of the biggest advantages of Apple Silicon to support multiple 16-lane PCIe GPUs Just. For. Me. Or keep supporting Mac OS on x86 Just For Me.

The biggest "advantages" of Apple Silicon aren't advantages to the market the Mac Pro has left to serve. AS as it currently is, an iPad SOC on steroids, is literally only disadvantageous to everything left over for a Mac Pro to be once you carve out the markets addressed by the other products in the range.

The Mac Pro's point is to be one machine whose design is Neutral, not welded to a pre-direction of what it's supposed to do (a failure contribution of the 2013 - prescribing it as a dual GPU OpenCL machine). So that it can be taken in whatever direction the user wants to take it - does this machine need a terabyte of ram to load databases, it can do that, does it need to drive 16 displays, it can do that, does it need ridiculous storage, it can do that.

That's literally the only reason to have a Mac Pro - to be the one truck in the product range, because they sure as hell don't have any trucks currently.

 

[theluggage]

The biggest "advantages" of Apple Silicon aren't advantages to the market the Mac Pro has left to serve.

I don't disagree.

The question is, what do you expect Apple to do about that? Develop an Apple Silicon CPU with 128 PCIe lanes and support for multiple TB of DDR5 RAM? For a machine that is never going to account for more than one percent of their sales, which would be useless in the other 99% of their machines and have no fundamental advantage over x86 boxes with the same discrete GPUs?

That's literally the only reason to have a Mac Pro - to be the one truck in the product range, because they sure as hell don't have any trucks currently.

...and, until about 2018, I'd have agreed with you, when a truck made sense to anybody with a dog or a jet ski. The trouble is, modern laptops and small-form factor systems can now (figuratively) drive across a field with a small-to-medium goat in the back, or tow and launch a 15' dinghy - so increasingly it's only people with 2 cows or a 40' yacht to tow who actually need a truck... and they're probably going to buy a boring old Toyota Landcruiser, not the Cybertruck that Apple would want to build to justify it's premium branding.

(I was going to say - "So what, Porsche don't make trucks!" but fortunately I checked, and apparently they do. Dang!)

 

[mattspace]

I don't disagree.

The question is, what do you expect Apple to do about that? Develop an Apple Silicon CPU with 128 PCIe lanes and support for multiple TB of DDR5 RAM? For a machine that is never going to account for more than one percent of their sales, which would be useless in the other 99% of their machines and have no fundamental advantage over x86 boxes with the same discrete GPUs?

I think that undersells the value of macOS, the fact that Apple thinks it can do what it's been doing to marginalise workstations is because they think people would rather buy hardware that doesn't fit with what they want, in order to keep their ecosystem integration.

How long that paradigm lasts in the face of regulatory aggressiveness with regards to Apple's hardware and software businesses being allowed to remain the same company, or tie their products to each other, is anyone's guess.

*edit* though I'd be surprised if Apple didn't have a processor-only modular chip in design, if only so they could build their own servers - while they may want to sell fixed appiances to consumers, for their own server rooms and deployments, I'd bet they spec for everything to be upgradable and configurable.

...and, until about 2018, I'd have agreed with you, when a truck made sense to anybody with a dog or a jet ski. The trouble is, modern laptops and small-form factor systems can now (figuratively) drive across a field with a small-to-medium goat in the back, or tow and launch a 15' dinghy - so increasingly it's only people with 2 cows or a 40' yacht to tow who actually need a truck... and they're probably going to buy a boring old Toyota Landcruiser, not the Cybertruck that Apple would want to build to justify it's premium branding.

I guess the real issue is that once you let people out of the walled garden, they tend to move everything out of the walled garden. If they can't have Safari on their workstation, they'll dump safari on their macbook, phone and ipad.

Very slowly to start, and then all of a sudden your integrative synergies turn into obstructive frustration.

 

[chanchowancho]

Mac Pro today is still very valuable for our company use case. We run two Mac Pros as Apple + Android App build machines.

The rack mount Apple Mac Pros are the only Macs able to be properly rackmounted (instead of just thrown on a shelf) in an enterprise data center. That alone was worth its weight in gold!

Previously ran a bank of Intel Mac Minis in the office, then an older Mac Studio crudely shelved in the datacenter, now the new rackmount Mac Pros couldn’t be easier!

 

[theluggage]

*edit* though I'd be surprised if Apple didn't have a processor-only modular chip in design, if only so they could build their own servers

I don't think they've been eating their own dog food for their own services for a long time. Even the 2019 Mac Pro was not a good server machine. There have been rumours of an AI server-optimised chip being in development but that sounds like it's going to exploit situations where having 512GB of unified RAM shared by an integrated CPU/GPU/NPU starts to beat the brute force of discrete GPUs with separate VRAM. I don't think it will address the former Mac Pro market.

I guess the real issue is that once you let people out of the walled garden, they tend to move everything out of the walled garden. If they can't have Safari on their workstation, they'll dump safari on their macbook, phone and ipad.

Not a problem as long as they still have a macbook. (iPhones and iPads can look after themselves).

The danger would be if a major third-party "pro" app stopped supporting the Mac because there was no flagship Mac Pro to target, then the lower-end Macs would lose it, too. Trouble is, Apple also needs them to optimise for Apple Silicon - pro software optimised for a discrete-AMD-GPU, external RAM Mac Pro isn't going to take advantage of the full power of an Apple Silicon laptop - and Apple Silicon-optimised software is what makes the MacBook etc. stand out from the crowd. The selling point of Apple Silicon is not what you can't do on a Mac Pro any more, it's what you can do on a MacBook or Mac Mini that you couldn't do before.

 

[mattspace]

Not a problem as long as they still have a macbook. (iPhones and iPads can look after themselves).

I keep coming back to the fact the (international) gamedev studio my partner works at was a major developer of Apple Arcade exclusives - like Apple made broadcast TV ads with actual actors in costume to promote their first major game, and in the entire studio, the only Apple device used in the *making* of the game, was a single headless Mac Mini build server in the server cabinet that all the windows workstation machines (all the artists are on Windows) used to compile the unreal project, before it was pushed to device (the iPhone plugged into the windows box).

The danger would be if a major third-party "pro" app stopped supporting the Mac because there was no flagship Mac Pro to target, then the lower-end Macs would lose it, too.

See above, I would be unsurprised if Unreal dropped macOS, given Unreal's push into tv and film production, which is all based on SteamVR and HTC MARS - all that VOLUME realtime LCD rear projection screen stuff, there's nothing for the Mac in that. Also, Apple's pushing Unity; a dying from mismanagement platform.

But another aspect of this is how many developers are abandoning "native" Mac software for cross platform janky UI libraries. If the polish, and quality that once made Mac software is gone, if everything's just going to a javascript driven UI contained in what amounts to a browser window, with no tear off palettes, etc - of they're going with QT6 (which frankly looks better than a lot of Catalyst or SwiftUI apps) or similar because that gets them the sort of rich UI libraries that Apple seems to be abandoning in their everything is a contained iPad screen, maybe there's less, and less to feel you're losing by going over to a Linux machine and setting up Plasma to have your familiar Mac menubar, etc.

Trouble is, Apple also needs them to optimise for Apple Silicon - pro software optimised for a discrete-AMD-GPU, external RAM Mac Pro isn't going to take advantage of the full power of an Apple Silicon laptop - and Apple Silicon-optimised software is what makes the MacBook etc. stand out from the crowd. The selling point of Apple Silicon is not what you can't do on a Mac Pro any more, it's what you can do on a MacBook or Mac Mini that you couldn't do before.

I was talking to a student dev this evening who is probably to to break huge in the Apple-centric world in the next 6-12 months because something he's working on is going to be massive in its implications for Apple devs - he uses a Macbook Air, but just bought a PC desktop for all the VMs he has to run as part of his dev environment. He's literally building an infrastructure that I suspect more third party devs making Apple apps will use than Apple's own and Apple kit and tools don't cut it for price / performance. It doesn't matter if an SOC is a rounding error faster than discreet memory and graphics, or uses less than $100 less electricity in a year, if it pulls the system out of affordability over a total lifespan it's missing table stakes.

I actually also don't really think the "power" of Apple Silicon is going to attract pro developers. I think what Apple Silicon's biggest, by an order of magnitude, effect is going to be, is flooding the Mac with shovelware iOS apps from the swamp that is the iOS app store. Low ambition iPad, and iPhone apps are going to be what the Mac ends up running for most people.

To take Apple's tools for example, the Apple Books app for the Mac used to have an entire publishing workflow built in, you'd plug in your iPad, work on your EPUB files on the Mac, and every time you saved an individual file, it would push that file to the proof version on the iPad, which would refresh and show you the change.

Then Apple replaced the Mac app with the en ified iPad one via Catalyst. The iPad app has no publishing workflow, so now there's no publishing workflow on the Mac, no incremental file by file updating on the Mac version, you have to manually erase the previous proof from the iPad, then airdrop the entire book across, then manually page through to where you want to check. And that often doesn't work because airdrop to Apple Books is janky, so your other option is to dump your epub file into the mac version, wait for it to upload to iCloud, wait for it to download from iCloud, and then preview.

Every single change made. I produce epubs up to 600mb in size. Apple isn't even maintaining pro tools for macOS. The iTunes Producer app has been replaced by a webpage tht doesn't let you change eBook preview screenshots.

The "you don't need a Mac Pro" Mac studio people will eventually be assaulted by the the "You don't need a Mac Studio" Macbook people.

 

[theluggage]

the only Apple device used in the *making* of the game, was a single headless Mac Mini build server in the server cabinet that all the windows workstation machines (all the artists are on Windows) used to compile the unreal project, before it was pushed to device (the iPhone plugged into the windows box).

...and if they couldn't do it that way, do you think they'd kit their artists out with Macs, or would they just not bother with Apple development? Note - they're not developing for Mac, they're developing for iOS which - unlike the Mac - has significant traction as a gaming platform. Apple are looking at iOS developers to help sell millions of iPhones and iPads - not a handful of "pro" Macs.

Unfortunately, the reality is that although Apple gained a commanding lead in pro graphics/video content creation back in the late 80s and early 90s - when Mac Hardware was head and shoulders more powerful than PC - they then spent the rest of the 90s losing that advantage as PC evolved and started offering cheap, commodity hardware with the required horsepower. Whether you blame Scully et. al, the Copeland fiasco, the Mac Performa - that ship had sailed, and the rest is slow decay.

If we're doing anecdotes - about 15 years ago I was working on a couple of video-related projects. In one case, at the London studio, all the editors were working furiously on white-box PC towers running Avid while the shelves were lined with dusty Mac Pros and Final Cut boxes, pretty obviously sitting there in case an old project had to be revived. Another producer was still using Macs - a MacBook Pro - and the payoff was that he could carry it around to clients offices and work with them on edits. At the time, the MBP was far better designed than most PC "flying bricks portable workstations" - trouble is, Apple were working with the same raw materials as their competitors - x86 and (pobably at that time) NVIDIA GPUs - so there was nothing to stop PCs playing catchup. Apple's Unique Selling Point in the video/graphics/3D market today is not that they offer the most horsepower-by-buck (they don't, by a long shot) but that Apple Silicon can do it adequately for many people's needs on an ultra-compact laptop.

I actually also don't really think the "power" of Apple Silicon is going to attract pro developers.

No. The only thing that will attract pro developers to Mac is a significant market of users who are attracted to Apple Silicon (because it makes nice thin laptops with good battery life & integrates nicely with their iDevices) along with low hurdles to supporting Mac (such as not having to switch to an all-Mac workflow).

But another aspect of this is how many developers are abandoning "native" Mac software for cross platform janky UI libraries. If the polish, and quality that once made Mac software is gone, if everything's just going to a javascript driven UI contained in what amounts to a browser window, with no tear off palettes, etc

...but the realistic alternative to that isn't better Mac software, it's less Mac software, or more browser-based apps, as developers don't want the effort of maintaining separate codebases for Mac, Windows, iOS, Android and (in some cases) webapp as well. That has changed from the good old days when Windows and Mac were the only things you needed to support.

I suspect that you'll find exactly the same in Windows development these days - with a lot more use of cross-platform stuff that can use the same codebase for Windows, Android, iOS and Web... because the last three are where the growth is these days.

He's literally building an infrastructure that I suspect more third party devs making Apple apps will use than Apple's own and Apple kit and tools don't cut it for price / performance.

If he's using PCs to run VMs then I guess he's using x86 VMs in which case I assume that it is targetting multiple platforms - see above. If it encourages developers to support Apple platforms its all good for Apple. You won't stop the drift to cross-platform tools. Or using modest laptops as "terminals" for cloud-based development.

You won't save the Mac just by forcing developers to buy them.

o take Apple's tools for example, the Apple Books app for the Mac used to have an entire publishing workflow built in

Was it nerfed before or after Apple Books abjectly failed to get any traction in a market already dominated by Amazon and B&N (and got whacked with a fine for price-fixing)? Beyond basic epub features, could it target features of Kindle readers/apps?

One has to assume that Apple are not completely stupid and when something like this gets axed it is because it wasn't paying its way.

 

[mattspace]

...and if they couldn't do it that way, do you think they'd kit their artists out with Macs, or would they just not bother with Apple development? Note - they're not developing for Mac, they're developing for iOS which - unlike the Mac - has significant traction as a gaming platform. Apple are looking at iOS developers to help sell millions of iPhones and iPads - not a handful of "pro" Macs.

They were developing for Apple Arcade, which is Mac, iPhone, iPad and AppleTV.

But no, they've moved on from Apple are are pivoting to console & PC, because the dirty secret of the games industry, is that no one working in it actually likes developing for mobile / low end devices. They don't like the resource constraints, the inability to actually show off what they're capable of on tiny screens, etc. You don't win awards (that are worth winning) or attract the best talent deploying on a potato platform. Just like you don't win friends with salad.

No. The only thing that will attract pro developers to Mac is a significant market of users who are attracted to Apple Silicon (because it makes nice thin laptops with good battery life & integrates nicely with their iDevices) along with low hurdles to supporting Mac (such as not having to switch to an all-Mac workflow).

I'm not sure that a maker of "Pro" software looks at the typical customer for a macbook, and sees a customer for pro software. More likely, they see someone who wants an iPad with better keyboard and mouse support.

...but the realistic alternative to that isn't better Mac software, it's less Mac software, or more browser-based apps, as developers don't want the effort of maintaining separate codebases for Mac, Windows, iOS, Android and (in some cases) webapp as well. That has changed from the good old days when Windows and Mac were the only things you needed to support.

Less Mac software is certainly what's happening. The design of macOS post Catalina is primarily about disguising that fact - the replacement of native style open/save as a menubar spawned sheets, to an easily-mimicked-with-CSS centrally floated box has a suspiciously timed effect of preventing webapps / electron / flutter / catalyst from being obviously non-native.

I suspect that you'll find exactly the same in Windows development these days - with a lot more use of cross-platform stuff that can use the same codebase for Windows, Android, iOS and Web... because the last three are where the growth is these days.

So everyone's servng and eating slop, and Apple is in the business of selling designer ceramics crockery and silverware cutlery to eat it.

You see where this leads?

You won't save the Mac just by forcing developers to buy them.

That's kindof what Apple does, though. Apple platform developers are increasingly the main customer use case / focus for the Mac.

Was it nerfed before or after Apple Books abjectly failed to get any traction in a market already dominated by Amazon and B&N (and got whacked with a fine for price-fixing)? Beyond basic epub features, could it target features of Kindle readers/apps?

No, it's been a long slow decline, that has little, if anything realistically to do with Apple being found guilty of masterminding a criminal conspiracy in the US District Court, whose only actual consequence was to require Apple to obey the law. It didn't constrain the company, or their eBook shop in any other way.

Mojave took away some javascript abilities, but Monterey was where Apple dumped the Mac app, and replaced it with the iPad app running in Catalyst.

Apple Books had plenty of traction, and was a vibrant platform providing the best reading experience, by a wide margin, on Apple devices. The problem was that it was managed by idiots. It's an absolute textbook case of the sort of self-sabotage that is at the core of Apple's DNA.

The ONLY problem Apple Books had, the reason they didn't do better against Kindle, was being fettered to the corpse that was having to buy Apple hardware to access the store or use the app. They literally had the lesson of history right there - iTunes and iPod on Windows (and later Apple Music for Android, AppleTV content being being available within every other streaming service, Apple TV device functionality being available on TVs directly).

Apple Books was a full Webkit implementation. It could do anything any web browser could do (it's efectively just two iFrames side by side with a novelty of being skinned to a draggable 3d page canvas). I had JQuery running in my books, recording and recalling book-internal individual preferences for element visibility, language, everything.

Kindle, and kindle readers were laughably primitive in comparison.

One has to assume that Apple are not completely stupid and when something like this gets axed it is because it wasn't paying its way.

It's more like someone deliberately cutting their fingers off, and then deciding that Tennis was never for them, and it was a waste of time for them to play it in the first place, because they aren't any good at it.

 

[Regulus67]

Apple Books had plenty of traction, and was a vibrant platform providing the best reading experience, by a wide margin, on Apple devices.

Thank you for explaining how Books used to be, and how it changed. Very interesting. I had no idea, as I am not a long time Apple user.

 

[257Loner]

The Mac Pro is at a crossroads of two of Apple's famous characteristics which are now opposed to one another:

1. Have the best design.
2. "Skate to where the puck is going, not where it has been."

One of the defining features seen in the history of the computer is miniaturization. There were actually ancient mechanical computers used to perform calculations, but the earliest electronic computers used vacuum tubes for logic circuitry and were housed in giant cabinets. Once the microchip was invented, computers began to shrink. Cabinet computers became desktop computers became laptop computers became pocket computers (smartphones) became wrist computers (smartwatches).

Despite the Mac Pro featuring a beautiful design, its place in the marketplace is not "to where the puck is going", for that is where smaller and more power-efficient computers like the Mac Mini and Mac Studio are today.

Sadly, once processors and computers are miniaturized even further, they will not take up enough physical space to deserve a case as pretty as the Mac Pro's. Thus will begin an era of retro computing where people will use computer boxes far larger than strictly necessary in order to enjoy a classic computing experience. Hey! If the keyboard and mouse are ergonomic enough, then count me in!

 

[123123123]

The "Steampunk" design may be a reference to the late, lamented Cheesegrater but it's really aimed at a different market that wants multiple high-end GPUs and biblical amounts of memory.

Not really relevant, but “biblical amounts of memory” has me laughing out loud.

In response to the OP’s question: I suspect that an answer is dependent on the unknown factor of how much macOS is optimized for the very specific range of CPU/GPU architectures Apple is churning out. For example, there’s not really any reason I can see that Apple couldn’t release a Mac Pro with slots to take upgrade cards loaded with an M-series processor and RAM – that’s probably a fairly simple hardware problem. However, how much effort would it take for Apple to make macOS use that efficiently? It’s been at least a decade or more since Apple released a multiprocessor system. The cost and effort to spool up the software to fit the hardware might be unpalatable.

The other issue is that the proportion of GPU-heavy workloads seems to be rapidly accelerating (pun intended). Is the current Apple Silicon architecture a good fit for scaling up GPU capacity alone? I’m hardly an expert, but I bet it’d take a lot of engineering effort. Not that it’s not feasible, or Apple couldn’t do it, but the RoI for such a niche product seems limited. (Also, cloud computing seems to be the direction things are going – why have a $10K Mac Pro if you can have a $3K Mac Studio and use cloud resources when necessary? Obviously that wouldn’t work for everyone, but it would for some, and that’d shrink the Mac Pro market even further.)

That being said, if Apple wants to make easy money on the Mac Pro, they can take the original cheesegrater case, put mounts for an ATX motherboard and PSU inside, and sell me that. IMHO, best case of all time. A lab I worked in had some, and they just look awesome.

 

[SDAVE]

I have a feeling Apple is going to completely kill off the Mac Pro line.

The only reason I see them keep it alive is:

1) Pros who need PCIe slots (audio + video/vfx folks)
2) Rack mounted servers
3) ML folks who need large memory

It's sad because the 2019 is such a beautiful machine. Quiet, powerful, great thermals, etc.

 

[mattspace]

Thank you for explaining how Books used to be, and how it changed. Very interesting. I had no idea, as I am not a long time Apple user.

That's just a cursory explanation - the sheer amount of footbulleting it took to ruin Apple's eBook ecosystem requires so much incompetence, for so long, that it's hard to ascribe it to incompetence, rather than a deliberate political play within the company.

For example, publisher relations / publisher support - the people you could speak to on the phone, 24/7 to discuss issues with how the store was working, or with how your books were appearing in the store, were not allowed to contact the engineering team (eg to report a bug in the way your book was rendering, etc) who made the iBooks / Apple Books application. While I was trying to chase down an EPUB issue with Books.app's rendering engine I had to be handed off from Books support, to Pro Apps Support, but they only dealt with Apple's authoring apps (iBooks Author), not handcoded XHTML EPUB.

Funny story, but that was where I found out that years before it was cancelled iBooks Author didn't actually have any staff assigned to maintaining or fixing it. There was one person in the entire company who did emergency bugfixes, but it wasn't part of his official job or duties. The team (mostly one guy) who actually made the app had long since disbanded / moved on.

From a user's perspective, there was an inflection point at which someone decided to juice the metrics of the store so they could get a bonus, and the app went from being user-library centric, to Apple-store centric. So whereas previously your primary interaction with the app was your library, and the store was a section you'd go to, that flipped to the store being the primary experience, and your library was a backwater you could visit.

It sortof mirrors the iTunes to Apple Music change, but there's no way to shut the store out of your face in Apple Books.

 

[Regulus67]

That's just a cursory explanation - the sheer amount of footbulleting it took to ruin Apple's eBook ecosystem requires so much incompetence, for so long, that it's hard to ascribe it to incompetence, rather than a deliberate political play within the company.
....
It sortof mirrors the iTunes to Apple Music change, but there's no way to shut the store out of your face in Apple Books.

The way you describe it makes me realise this is part of why I am not keen on the latest mac hardware.

There is no software support on Apple hardware for older OS versions, than it shipped with.
So in my case, the M1 and possibly M2 generation hardware will be the last hardware from Apple that I will consider.
This was not an issue I ran into back when I used PCs. I could get new hardware, but still use an older OS version.

I have one iPhone 15 Pro Max that I only use rarely for video recording. And I have not updated the OS, and probably never will. Ironical really, as IOS updates was what made me get my iPhone 6s (still my main phone), and switch to the Apple ecosystem.

No wonder I love the iMac Pro and the Mac Pro 7,1. The machines are good enough for my use. And let me avoid part of the AI development, as I am increasingly alarmed by it.

 

[theluggage]

For example, there’s not really any reason I can see that Apple couldn’t release a Mac Pro with slots to take upgrade cards loaded with an M-series processor and RAM – that’s probably a fairly simple hardware problem. However, how much effort would it take for Apple to make macOS use that efficiently? It’s been at least a decade or more since Apple released a multiprocessor system. The cost and effort to spool up the software to fit the hardware might be unpalatable.

Yup, and I'm pretty sure I (amongst many others) mentioned that possibility during the speculation back before the 2023 Mac Pro came out. It could have expanded on the MPX slot idea in the 2019 Mac Pro to connect power, Thunderbolt etc. to the modules.

Another possibiliy would be something like a 1U rackmount version of the Mac Studio Ultra (maybe 2 to a unit) with 6-8 Thunderbolt ports on the back - fill a rack with these and network them together with TB and you'd be cooking with gas (given software that could use it...)

You're not talking about a multiprocessor system in the sense of the old Mac Pros with dual Xeons (sharing the same RAM, GPU etc.) which work in a fairly similar way to multi-core processors, esp. from the point of view of applications. What you're talking about is more of a cluster system of complete computers communicating by some sort of fast link (maybe PCIe, Thunderbolt or some other super-fast interconnect).

Computer cluster - Wikipedia

en.wikipedia.org

...that is very much a thing, as you see from the Wiki article, and the basis for many modern supercomputers. Adding support for one of the various standards to Mac OS would take work, but would hardly be a Manhattan Project given the support already exists in Linux. The question is, what does MacOS and/or Apple Silicon have to bring to this arena - in which Linux and ARM are already well established?

Here's what's just come online in the UK:

UK's most powerful supercomputer Isambard-AI comes online

The Isambard-AI supercomputer is made fully operational as the government unveils fresh AI plans.

www.bbc.co.uk

One interesting thing about that is it is using Grace/Hopper processors which have several similarities to Apple Silicon Ultra processors - an ARM-based CPU with on-package LPDDR RAM and a fast chip-to-chip link analogous to Apple's UltraFusion connecting two dies together (it doesn't have to be used for such insanely large clusters but its still a serious-callers-only product). In this case, though, one die is a CPU and the other a GPU - you couldn't really get there by gluing together Mx Max chips (still not clear how Ultrafusion would work physically with more than 2 dies). Apple probably could do something like it with ASi tech, but it would be a whole new die.

So there are plenty of possibilities for Apple to do something interesting with high-end Apple Silicon systems, the question is whether there's a market niche for Apple to exploit which would justify the investment and risk - and the results would be something different that would not really address the needs of much of the previous Mac Pro market, nor would they necessarily be viable consumer/prosumer retail products.

Even the 2019 Mac Pro was pushing a "prosumer" product into the price range of specialist, bespoke systems - where building something unique from discrete components was more viable.