Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.


Apple today updated its Mac Pro desktop tower with the all-new M2 Ultra chip, which features a 24-core CPU, up to a 76-core GPU, and support for up to 192GB of memory. Apple says the new Mac Pro is up to 3x faster than the fastest Intel-based Mac Pro.

Mac-Pro-Feature-Teal.jpg

The new Mac Pro has the same overall design as the previous Intel-based model. However, the tower is now equipped with eight Thunderbolt 4 ports instead of four, two higher-bandwidth HDMI ports that support up to an 8K display at 240Hz, seven PCIe expansion slots, dual 10 Gigabit Ethernet ports, and a headphone jack that supports impedance headphones. The computer can be configured with up to 8TB of SSD storage.

Like the new Mac Studio, and the latest MacBook Pro and Mac mini models, the Mac Pro now supports faster Wi-Fi 6E and Bluetooth 5.3.

The new Mac Pro is available to order starting today, and it will begin arriving to customers and launch in stores on Tuesday, June 13. In the U.S., pricing starts at $6,999, whereas the previous Intel-based Mac Pro started at $5,999. A rack-mounted version of the Mac Pro remains available from $7,499.

Article Link: Apple Unveils New Mac Pro With M2 Ultra Chip, Extra Ports, and More
This Mac starts at £7,199. What the hell? How did the starting price for a Mac Pro get to this level of lunacy?
 
Just go to hp.com or dell.com and see what they offer (cost and features)...
You'll reconsider your Mac Pro concept.
Oh I did and they support way more options with superior CPU and GPU. Mac Pro 2019 already supports up to 4 GPU so why do you even bother to justify Mac Pro 2023's pathetic specs?
 
192GB max memory limit is incredibly small. Real workstations need terabytes of RAM. It's basically an overpriced Mac Studio with PCIe for IO, which can normally be done via Thunderbolt these days.

Really surprised Apple went with this kind of architecture. They needed a blowout product.

Agreed. After all that buildup they just did the most obvious thing. It’s not really a proper replacement.
 
  • Like
Reactions: George Dawes
you probably won't find a machine that can give full dedicated bandwidth to all of its slots at the same time

We run EPYC Milan based servers, with 7x 16-lane slots, each with full, uncontended 16 lanes of bandwidth direct to the CPU, no PCIe switch chips.

Two of our (storage) servers come pretty close to saturating all of those lanes, being loaded with NVMe HBA cards bifurcating each 16-lane slot into 4x 4-lane NVMe interfaces for multiple U.2 SSDs.

It would be a crying shame, yet seemingly ominously inevitable, if the ASi Mac Pro's slots and TB4 ports are contended in some way...
 
I wonder if there are any M.2 or U.2 slots? I would love to be able add additional *internal* storage without necessarily taking up a PCIe slot.
One takes up PCIe lanes with M.2 or U.2 slots only in a way that they can only be used for that. Adding PCIe slots allows the most flexibility.
 
With NVLink spanning memory load was somewhat possible (e.g., two linked 24GB VRAM GPU cards could manage 48GB of data). However, it was never widely adopted — presumably why SLI and NVLink are now EOL. Multi-GPUs (i.e., cards) are still plenty beneficial nowadays but it’s more of working in parallel (e.g., each renders a different frame or each processes a different simulation).
Exactly. There’s a lot of use cases where working on smaller chunks of data in parallel pools of memory is advantageous. And, there are use cases where having ALL the data in one big chunk to operate on is the only way to go. For the second case, no one makes a GPU with as much RAM available as Apple Silicon.
 
  • Like
Reactions: bcortens
No probably not, but I also don’t believe that the new M2 ultra is as fast as let’s say if I would add the latest mpx modules available into my Mac Pro.
Good point, it would depend on what you’re doing, of course. I’ve always felt that if what someone has works for them, then stick with it until it doesn’t!
 
Where did you read for a fact that the new MP does not support external gpus?
It’s not referenced anywhere, but I’ve frequently been told the chip doesn’t support them, either internally or via e-GPU. The list of cards on Apple’s website are all audio, video or IO. And the slots are no longer MPX compatible. Plus it’s Apple; ****ty old stupid ‘we don’t play nice with anyone’ Apple. I’d absolutely love to be wrong, but I’ve seen no evidence to the contrary.

EDIT: There you go… https://appleinsider.com/articles/2... Silicon has finally arrived,or a 76-core GPU.
 
  • Like
Reactions: CarAnalogy
It’s not referenced anywhere, but I’ve frequently been told the chip doesn’t support them, either internally or via e-GPU. The list of cards on Apple’s website are all audio, video or IO. And the slots are no longer MPX compatible. Plus it’s Apple; ****ty old stupid ‘we don’t play nice with anyone’ Apple. I’d absolutely love to be wrong, but I’ve seen no evidence to the contrary.
I disagree with the reasoning behind this. They can play nice when they want, instead this is the unified memory architecture in action. By getting all of their products on a unified memory platform developers can really optimize targeting the apple silicon paradigm instead of having to keep around old assumptions about memory copies and locality. I think long term this has a lot of potential for improved performance.
 
  • Like
Reactions: AdamBuker
Where did you read for a fact that the new MP does not support external gpus?
Apple’s developer documentation does not define a way for a developer to access to ANY GPU other than the one on the SoC. It’s been that way since the first Apple Silicon developer videos were made available. And, with each year, they continue to define additional Apple Only GPU methods. So, currently, MP can’t support external or internal GPU’s.

COULD Apple flip and do something completely different? Absolutely. By that token, they COULD sell Mac Pro’s for 14 cents. Both things, based on Apple’s developer documentation, are just as unlikely.
 
  • Like
Reactions: skippermonkey
I disagree with the reasoning behind this. They can play nice when they want, instead this is the unified memory architecture in action. By getting all of their products on a unified memory platform developers can really optimize targeting the apple silicon paradigm instead of having to keep around old assumptions about memory copies and locality. I think long term this has a lot of potential for improved performance.
Maybe long-term, but currently the situation sucks.
 
this is the unified memory architecture in action. By getting all of their products on a unified memory platform developers can really optimize targeting the apple silicon paradigm instead of having to keep around old assumptions about memory copies and locality. I think long term this has a lot of potential for improved performance.
And it’s already shown off some of its potential as code written for macOS Intel machines run FASTER on macOS Apple Silicon machines (using Rosetta2). It’s a good solution that the wider market hasn’t had an opportunity to even evaluate due to the preeminent architecture of the day requiring the highest performance GPU’s to be separated from the highest performance CPU’s by PCIe. People who say they’d love to see evidence to the contrary has likely already seen evidence to the contrary, but they didn’t like the evidence that was to the contrary because, well, it was to the contrary. So, they’re still waiting for the evidence :)
 
  • Like
Reactions: bcortens
And it’s already shown off some of its potential as code written for macOS Intel machines run FASTER on macOS Apple Silicon machines (using Rosetta2). It’s a good solution that the wider market hasn’t had an opportunity to even evaluate due to the preeminent architecture of the day requiring the highest performance GPU’s to be separated from the highest performance CPU’s by PCIe. People who say they’d love to see evidence to the contrary has likely already seen evidence to the contrary, but they didn’t like the evidence that was to the contrary because, well, it was to the contrary. So, they’re still waiting for the evidence :)
By people, you mean me? I said I haven’t seen evidence to the contrary that Apple silicon supports external GPUs. And then I provided the evidence. So your snarky comment makes zero sense on many levels.
 
External GPU support is coming for ASi, Asahi Linux ensures this will happen one way or another.

IMO, the big problem for Apple here is the massive distraction from Apple Vision Pro. The drive by introduction of the Mac Pro was frankly embarassing and Apple will loose legions of fanatics that sell Apple products because those fanatics influence buying decisions of a large number of people they know.

The fact that Apple missed the M3 launch (R1 caused its demise) suggests they have left the door open a crack to emergence of a competitor in their space. For a while now it has been apparent that Qualcomm was being driven into oblivion. After yesterdays WWDC, Qualcomm may have a breath of new life if they can take advantage of Apple’s distraction not to mention Samsung and nVIDIA. And if nVIDIA gets their ARM acquisition back on track it is a whole new ballgame. Competition combined with the increased government scrutiny of Big Tech suggest the forces building against Apple are becoming formidable.

Apple sales will drop precipitously across all of their business segments.
 
Last edited:
  • Like
Reactions: George Dawes
The fact that Apple missed the M3 launch suggests they have left the door open a crack to emergence of a competitor in their space.
Apple has a monopoly on macOS systems, they’re the only company that can legally make them. There IS no competition in that space.
 
If I was on the market for a Mac Pro, I would be pretty disappointed.
Funny, I am in the market for a Mac Pro and I am very happy about it. A fully configured Mac Pro is only $11,800, much less than we paid for the 96GB in our 2019 Mac Pro.
Luckily the Mac Studio is good enough for me.
Great that Apple has customers with different needs and therefore makes different products with different specs.
 
  • Like
Reactions: CWallace
Funny, I am in the market for a Mac Pro and I am very happy about it. A fully configured Mac Pro is only $11,800, much less than we paid for the 96GB in our 2019 Mac Pro.

Great that Apple has customers with different needs and therefore makes different products with different specs.
I was waiting for a new MacStudio release to upgrade an ageing MacPro (2013) was kind of hoping for a MacPro but doesn't seem to justify the difference between both for my needs. Anyway one looks at it this lMP and can't help to think is a half baked MacPro.Apple could have saved MP for October/ end of the year and ship with an M3 as it is looks almost as if for some reason most people aren't aware Apple was in a hurry to ditch the last link to Intel.
If there was only a Studio release looked like they were still innovating, releasing MP as it is looks like: we tried but this is the best we can do.

Edit: BTW, still on 2013 MacPro because it was at least RAM upgradeable, at the moment of purchase it was painfully expensive but in due time changed from 16 to 128GB RAM
 
Last edited:
he fact that Apple missed the M3 launch (R1 caused its demise)
What makes you think R1 caused its demise and not TSMC's delays in getting 3nm production ready?

On a side note, this is from a post by 512 Pixels (take it for what it's worth):

"Several of us who cover Apple have heard that there are those inside the company that did not want this machine to see the light of day, believing the Mac Studio to be enough to hold down the high-end of the Mac line. Seeing the machine that Apple announced this week, I think they may eventually get their way."

 
  • Like
Reactions: chfilm
What makes you think R1 caused its demise and not TSMC's delays in getting 3nm production ready?
I recall some reporting of internal concern that resources would be spread the design team too thin. At the time I didn't understand how Vision Pro would concern the A/M series design teams but the existence of the R1 does now confirm this was going on.

"Several of us who cover Apple have heard that there are those inside the company that did not want this machine to see the light of day, believing the Mac Studio to be enough to hold down the high-end of the Mac line. Seeing the machine that Apple announced this week, I think they may eventually get their way."
Indeed, this is likely true. Are there any benchmarks showing Mac Pro vs Mac Studio yet? I wonder if the Mac Pro is running the M2 Ultra at a higher clock because it has a better thermal envelope?

I think there might be some pre-existing IP when it comes to Vision Pro ...

 
Apple has a monopoly on macOS systems, they’re the only company that can legally make them. There IS no competition in that space.
With the emergence of Asahi Linux though, Apple no longer has a monopoly on which OS can be run on the ASi hardware. Thus, a competitive ARM processor could emerge that would enable users to run whatever they want. And, the way regulators are heading, it may not be long before Apple will be compelled to allow macOS to run on non-Apple hardware. This is where the crack is. If a strong competitor arises that is persuasive with the regulator, Apple's ecosystem could take a direct hit.
 
And, the way regulators are heading, it may not be long before Apple will be compelled to allow macOS to run on non-Apple hardware.

macOS makes up far too little of the overall PC marketspace in any country to ever be considered an anti-competitive monopoly position.

I mean they couldn't make the case against Windows of all things.
 
  • Love
Reactions: smulji
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.