MP All Models Apple sleeping on the mac pro is comically bad timing

[HuskerNKS]

Oh my goodness LARP harder 🤣

Showing my age here. I had to look up LARP... 🤣

I would have just went with...

 

[Charlie Bonesx]

Hash cracking. Any FP64. Even a cluster can't touch any Mac Pro except the 2023. Y'all really need to stop trying to tell us what we need. 😂

so apple should be be building better tools for organized crime? 🤣

 

[Charlie Bonesx]

There's a big enough population of us that felt the pain when Xserve, the RAID, server app, Time Machine and all the networking products were discontinued. If they were to introduce a quiet 1U server with unlimited everything, it would sell like hotcakes.

Time Machine has not been discontinued

Yup, a turnkey APFS / iSCSI NAS, that can be an iCloud cache / alternative (Remember Transporter Sync?), but you know, subscriptions and services are Timmy's legacy.

APFS doesn't work like that

 

[theluggage]

There’s no reason why Apple wouldn’t reintroduce a tower Mac in the future.

...apart from all the reasons that have been repeatedly pointed out in this thread - and many other threads since the dawn of Apple Silicon...

TLDNR: huge investment in a new PCIe/DDR-centric die to produce a "me too" tower for a shrinking market - a die that throws away most of the advantages of Apple Silicon and will at best perform as well as whatever AMD GPUs it is driving, which would work just as well in a cheaper, and more tailor-able, Threadripper box. Plus, most people want NVIDIA for GPU compute, anyway.

In 2017, Macs were still, basically, just PCs licensed to run MacOS and Apple could have picked up the phone to China and ordered a few container-loads of "xMacs". Even the 2019 Mac Pro owed most of its power to the new Xeon-W chips from Intel and AMD's latest GPUs.

Apple Silicon has changed that - if you want a PCIe tower then the existing Max/Ultra dies are simply not the tool for the job. That would be a problem if Apple Silicon had flopped in mobiles, laptops and SFF desktops - but it was a success, it's put most of Apple's range ahead of the game and left the PC industry playing catchup. The Mac Pro was the price to pay for that.

 

[Flint Ironstag]

so apple should be be building better tools for organized crime? 🤣

Your complete ignorance is showing if you think cryptanalysis and accurate math are the sole province of organized crime.

Time Machine has not been discontinued

Yeah, Time Capsule. You managed to get that right.

 

[Omega Mac]

...apart from all the reasons that have been repeatedly pointed out in this thread - and many other threads since the dawn of Apple Silicon...

TLDNR: huge investment in a new PCIe/DDR-centric die to produce a "me too" tower for a shrinking market - a die that throws away most of the advantages of Apple Silicon and will at best perform as well as whatever AMD GPUs it is driving, which would work just as well in a cheaper, and more tailor-able, Threadripper box. Plus, most people want NVIDIA for GPU compute, anyway.

In 2017, Macs were still, basically, just PCs licensed to run MacOS and Apple could have picked up the phone to China and ordered a few container-loads of "xMacs". Even the 2019 Mac Pro owed most of its power to the new Xeon-W chips from Intel and AMD's latest GPUs.

Apple Silicon has changed that - if you want a PCIe tower then the existing Max/Ultra dies are simply not the tool for the job. That would be a problem if Apple Silicon had flopped in mobiles, laptops and SFF desktops - but it was a success, it's put most of Apple's range ahead of the game and left the PC industry playing catchup. The Mac Pro was the price to pay for that.

I'm telling ya M-Blades in a tower, shed, trailer... warehouse/building to the sky!

- Mac Studio Pro (4/6/8/10 Blade bay)
- w/ M-Blades Ultra

We're back at buildings as computers, or that's alway been part of the dev loop, how long before you have todays data centre on your desk? Anyone.. anyone.. Apple.. APPLE??? 😉

 

[Charlie Bonesx]

Your complete ignorance is showing if you think cryptanalysis and accurate math are the sole province of organized crime.

ok I'll bite.

what are you hash cracking for?

 

[mattspace]

APFS doesn't work like that

The point being that it mounts on a Mac transparently, as if it were a local APFS volume, and so works like a local drive. Rather than SMB, which can't do certain things a local disk can do.

 

[Charlie Bonesx]

The point being that it mounts on a Mac transparently, as if it were a local APFS volume, and so works like a local drive. Rather than SMB, which can't do certain things a local disk can do.

I guess I just don't see the business case for apple to sell a NAS device?

 

[bzgnyc2]

Were there ever Mac systems with slotted VRAM?

Effectively the Mac Mini 2018's VRAM is slotted since those basically use system RAM for VRAM and its RAM is upgradable.

And while the VRAM on the RADEON cards of the Mac Pro aren't upgradable, one could replace or add cards without throwing the whole system away.

The larger point is that Apple backed itself into a corner with its design tradeoffs. Yes there are (substantial) upsides as we've all seen -- and they work great in the laptop/Mini market -- but now we're also seeing the downsides that people warned about. Buyers' choices for full desktop (Studio+ segment) are either buying a computer now with 96GB for the life of the computer or nothing. Max of 1.5TB RAM + 2x64GB VRAM/HBM2 on the 2019 Mac Pro down to 96GB max is quite a drop...usually things don't go backwards after 7 years.

 

[Charlie Bonesx]

down to 96GB max is quite a drop..

don't you mean 256GB max?

doesn't the largest NVIDa GPU max out at 192 GB?

 

[bzgnyc2]

Apple abandoned the "high-end" in 2012/2013 when they let the original Mac Pro wither on the vine and replaced it with the Trashcan (...which, I'm pretty sure, they then planned to replace with the iMac Pro in 2017).

Every Mac Pro since 2012 has been a "one and done" - the Trashcan, the iMac Pro, the 2019 MP (which priced out many potential buyers), the 2023 MP (same case - totally different product). I don't think Apple are particularly "strategic" about the high end).

The Mac Studio is clearly where Apple wanted to go with the Trashcan - and a much more credible product - and has been updated twice which hasn't happened to a Mac Pro since about 2012....

Meanwhile, Apple Silicon has proven to be a major strategic advantage at the low/mid-range all the way up to the Studio. I think the "high end" is something they're happy to sacrifice - they don't need to have a foot in every pie.

Agree, abandoning the "high-end may have been part of Apple's business strategy and it may have made the most sense overall. I think the OP was arguing it was a mistake. Time will tell. However, let's not pretend their current designs are not without tradeoffs or there were no alternatives. Maybe the best trade-offs given the chosen business strategy that may work out quite well for them but not without trade-off. And nobody likes being the traded-off.

 

[bzgnyc2]

don't you mean 256GB max?

Those no longer exist for purchase. The most RAM you can order in a Studio now is a 96GB (with an M3 Ultra rather than M4 Max or better processor).

Yes Apple had Studios that went up to 256GB and 512GB not too long ago but they had to drop for reasons of their supply chain. I get the supply chain issues but that doesn't change the situation for the buyer. Unlike other designs, there's no 3rd party or even future Apple upgrade option.

 

[leifp]

GPU expansion is incompatible with unified memory. Also, most of the people begging for dGPU support also seem to be begging for NVIDIA/CUDA which ain't gonna happen if only "because politics".

RAM expansion could still be done with unified memory but that loses the bandwidth advantages of having the RAM mounted directly on the package with short-as-possible traces. Also, when Apple Silicon started, LPDDR RAM had to be soldered direct to the motherboard. There are now press-fit LPCAMM modules - but they don't seem to have taken off (last I looked I couldn't actually find them for sale). Using DDR RAM as cache might be an idea - but doesn't have the advantage of working like VRAM and is subject to the same price rises and shortages as any other RAM.


The Mx Ultra has 32 PCIe lanes (22 free for slots, even that means sharing some with internal peripherals via a switch) - which is only acceptable bcause it doesn't rely on external GPUs. Competing Xeon/Threadripper CPUs have 128 lanes.

Matching that would need Apple to design, produce and tool-up for a new Apple Silicon die just for the relatively tiny Mac Pro market, which would be extortionately expensive - even the current Mx Ultra relies on the economy of using two fused Mx Max dies (as used in MacBook Pros).

You'd end up with a so-so ARM-based tower with a massive "Apple tax" but with no huge advantages over generic PC hardware, limited software compatibility and no "trickle down" of Apple Silicon/Metal-optimised software for Mac small-form-factor, laptop and tablet systems.



A Mac Pro with AMD and NVIDIA GPUs (and a hugely expensive custom CPU with enough PCIe to drove them) would turn in similar performance to a Threadripper with exactly the same AMD/NVIDIA GPUs - and that's assuming Apple's drivers were on a par with the Windows/Linux ones. Apple's own GPUs still wouldn't have the "unified RAM" advantage.

I wouldn't rule out Apple producing some sort of "compute module" with an Mx Ultra SoC and PCIe for fast clustering, but that would be vastly different from any Mac Pro that ever existed.

Those are constraints due to design choices not actually impossible things. They can be designed against. They haven’t been. That is the entirety of the concern and I wish they had and they still could but they presently don’t.

 

[leifp]

You would? How much? Note that these kinds of machines exist — they are Nvidia's datacenter ML solutions, and their cost far exceeds that of regular workstations. Fast switches, large chips, and a lot of RAM are not cheap.

There are significant engineering challenges to making what you envision possible. For example, Apple chips don't have enough high-speed I/O to drive high-performance external GPUs, because unified memory controller take a lot of die edge space (and that's where the I/O is routed from). Traditional server CPUs don't have this problem because they don't use high-bandwidth RAM. Now, this is potentially solvable, of course (and Apple has filed relevant patents!) but the solution would be prohibitively expensive.

Looking at this realistically, your proposal boils down to one of the following options:

- create a server-grade Apple Silicon system with more performance and RAM capabilities (which would drive R&D and manufacturing costs to a point that most customers likely won't find attractive)
- create a dedicated expandable system that can be equipped with third-party solutions (which would break Apple's fundamental programming model and fragment the software ecosystem for the sake of a niche market)
- create a separate line of computers targeting workstation market and potentially ship them with Windows/Linux

Overall, it doesn't seem likely that there is much business sense for them going that route. In fact, Apple's own custom datacenter use modular systems that leverage multiple "small" managed SoC boards rather than large systems like you describe. It just appears to be a more economical solution. And for the end customer interested in local inference, the Studio with Ultra chips offers plenty of RAM and compute to power many relevant applications. The M5 Max is already equivalent to Nvidia Spark for half-precision work (Nvidia of course is massively ahead for fixed-precision and quantized matrix compute). Future systems will bring new capabilities.

That’s beyond my pay grade. I don’t see why that level would be required for what I propose but I’m not a hardware guy (as in: those who actually design/build chips and motherboards, etc) so I may be missing something (or many things). However, the core from my understanding would be similar to what desktop setups with iGPUs and dGPUs have been doing for quite some time. The RAM pool and PCIe channels are the constraints. Back in the Northbridge and Southbridge days these things were already handled, if insufficiently for my proposals.

 

[bzgnyc2]

How is that going to help you with AI? The benefit of unified memory for AI is that it's accessible by the GPU.

Not everything is AI. But even with AI, paging data between main RAM and GPU/VRAM is better than nothing. Certainly better than swapping to disk. Having a giant block of RAM (e.g. VRAM) directly accessible to the GPU is of course ideal but then the question is what is your fallback when you have more data than will fit?

 

[theluggage]

The point being that it mounts on a Mac transparently, as if it were a local APFS volume, and so works like a local drive. Rather than SMB, which can't do certain things a local disk can do.

...but that doesn't need any Apple-specific features on the server. The server presents an iSCSI device, the Mac sees that as a SCSI block device (the whole point of iSCSI), which it should be able to format as APFS (and create a Time Machine volume if that's what you want).

It's certainly worked at some point: https://community.synology.com/enu/forum/17/post/109738

If APFS doesn't properly support iSCSI for some reason that's a MacOS problem on the client machine. There's nothing an Apple-made server/NAS could add to that over a generic iSCSI device - other than offering some non-standard proprietary alternative to iSCSI - which would be a bad thing, because people need Macs to work well in mixed environments.

(Apparently NVMe over TCP is a thing too...)

 

[BSDnostalgia]

And while the VRAM on the RADEON cards of the Mac Pro aren't upgradable, one could replace or add cards without throwing the whole system away.

Sorry, I'm seeing the other side of that trade. I think Nvidia and Apple are both on the same track: ARM CPUs with unified memory GPUs attached. 48GB VRAM not necessary because that is the system memory now. Goodbye x86 and PCIe. Nvidia is going with Linux and Windows. Apple with MacOS. Obviously I prefer MacOS today, but, Linux could work for gamers when Nvidia starts selling its desktop AI systems as gaming systems, which I assume that they will. What I like about Apple is the low-cost (shortly, not today, but downward compatible) TB5 connections for consumers, but, what I like about Nvidia is the higher-speed but not terribly expensive QSFP interconnect. e.g. LinkX
Apple would do well to support QSFP LAN links on its next Studio Ultra. And, would do well to facilitate gamers better.**

** Not to try to cover everything, but, it appears that Apple's GPU architecture scales almost as well as Nvidia, except for Nvidia's current secret sauce, Ray Tracing.

 

[Bug-Creator]

Hash cracking. Any FP64. Even a cluster can't touch any Mac Pro except the 2023. Y'all really need to stop trying to tell us what we need. 😂

So how many 10000 are planning to do that under macOS? Doing a bit of market research for Apple here 😛

But it seems you asking just for more compute and maybe RAM which Apple could do just as well with an M6-Über MacStudio (maybe with a bigger case).

Also no ones cares what you need, no ones cares about what we might or might not be telling you on what you need.

Apple cares about what the can sell in enough quantity and at the right price to make a good profit over development and production costs which at the moment seems stop with the M3 Ultra Studio.

 

[mattspace]

I guess I just don't see the business case for apple to sell a NAS device?

Onboard flash storage is getting more and more expensive, macOS is getting bigger and bigger, with more and more shovelware apps included as a part of the immutable system, and more and more of the systems media-heavy file storage is either required to be on th boot volume, or on an APFS external volume that only one computer can plug into at a time.

The business case is "your entire family photo / media library, but kept in your house", which also does network time machine, and smart home control hub, and can then back itself up to a big, expensive Time Machine for Apple Home offsite backup, with a big fat storage subscription.

It's the hardware incarnation of the Apple family plan / family AppleID.

 

[theluggage]

Those are constraints due to design choices not actually impossible things.

In the case of unified memory, no, it's not a design choice - if the GPU, NPU and CPU aren't on the same die (or some sort of ultra-fusion package) as the memory controller then it isn't unified memory - at least not the sort of "unified memory" that Apple Silicon boasts. Let alone if the GPU/NPU are plugged into a PCIe bus - which is what people have been discussing in this thread.

You want to imagine some sort of fast - but still plug-in - GPU-CPU interconnect that also shares RAM and runs at comparable rates to the "fabric" on a SoC then go ahead and move the goalposts - it's nothing we've seen in a Mac Pro before.

Using DDR5 DIMMS might be a "design choice" but it still probably involves a whole new die and packaging technique just for a Mac Pro that accounts for a tiny fraction of Mac sales. Especially if you also want 2019 MP-like RAM capacities (which, BTW required Intel's M-suffix chips that supported 2TB and added thousands of bucks to the higher-end Xeon options on the MP).

Of course, everything is a design choice... Apple could have made a "design choice" to build the Mac Pro around an AMD Threadripper, or a "design choice" to release a range of plush garden gnomes instead...

Sure, there was a "design choice" but the choice was to use the same Apple Silicon dies for which the development/tooling costs could be shared with the far bigger-selling MBP and Studio & which punched above its weight because of unified memory.

 

[jakey rolling]

Depends. The internet runs on *nix you know.

Except for some (sometimes rather large) islands of the internet that don't. Many of those islands running Windows and Windows-related technologies (Hyper-V, Azure) are where some pretty large enterprise applications live and are also where AI agents are starting to actually become useful.

CEOs think that ordinary people use Windows, but, plenty of ordinary people use Macs. A few services still use BSD Unix directly (e.g. FreeBSD). All *nix. And, plenty of expertise in administering *nix out there, including in most of Nvidia's institutional customers.

Thinking CEOs are a monolith of technically stupid people: You read a few too many Dilbert comics, my dude. Probably the second-largest group of Mac users in the enterprise space (behind advertising and media departments) are C-Suite execs. Some larger enterprises might still be administering *nix systems in-house, but most SMBs are moving into the cloud, where the bulk of *nix administration is being done by cloud providers.

You have to be joking. Only CEOs think it is cheaper to operate and manage Windows. The advantage of Windows is that it is familiar to managers in large companies and their minions. That, and, Microsoft's iron grip on Excel. Those are big advantages, but, could be overcome if these companies were actually as cost-conscious as they say they are.

CEOs are always cost conscious. If there is a measurable cost savings that can be demonstrated to a CEO of moving to a non-Windows platform, it should be quite easy to convince most of them to move. The fact that we're now 30+ years into the internet age with *nix "running" things, there should be a lot of good data out there to build a proposition. Excel is an application minions use. CEOs and other C-Suite exects don't use Excel for anything more than consuming pre-made content provided by those minions. Excel isn't keeping enterprises on Windows at the server level. The things keeping them on Windows is simply that people like to say "running Linux is cheaper" without actually having any numbers to back that assertion up.

 

[iphonenine]

The Mac Pro supported up to 2000GB/2TB RAM + massive high speed 16 TB storage + multiple high-end GPUs and plenty of space for more hardware. It costed thousands of dollars, it was not affordable at all compared to other stationary Macs like the mini and Studio, so Apple discontinued it. But at the same time, it was not built for AI: The M5+ MacBook Pro are more than powerful enough for AI. The Mac Pro was meant for power user professional workflows like 8K pro res photo and video editing, game development, engineering software, etc. Again for non-AI tasks.

 

[theluggage]

The business case is "your entire family photo / media library, but kept in your house", which also does network time machine, and smart home control hub, and can then back itself up to a big, expensive Time Machine for Apple Home offsite backup, with a big fat storage subscription.

Now explain the business case for Apple making their own NAS... when Synology, QNAP, Western Digital and a dozen other suppliers make Mac-compatible NASs and "home cloud" boxes that work fine with Apple file sharing and Time Machine. Or, when you can just use a Mac Mini and a neat-looking stackable SSD.

 

[Charlie Bonesx]

"your entire family photo / media library, but kept in your house", which also does network time machine,

I do this with an old Mac mini