The real question is will Apple produce a high-end version of this Mac Pro? Right now, the only available version is entry-level, with silicon that overlaps with the high-end Mac Studio. There is no high-end Mac Pro.
The silicon would have to go beyond the Ultra. Beyond the beyond (ultra means "beyond"). Apple has said enough to make it clear that they did design such a tier, but it's not clear what that was. All that is known is that it was aborted and its code name might have been Jade 4C-Die.
But there are some things that we can surmise:
#1 is that killing Jade 4C-Die wasn't a financial decision, it was about performance. It didn't give them what they were looking for at the high-end Mac Pro starting price of $14K (i.e., double the entry-level price, like the Mac Studio). My assumption is that the reality of the fundamental limits in M1/M2 graphics performance meant its time had not come.
#2 is that Unified Memory is here to stay. Whatever Jade 4C-Die was, and whatever it will become, it will be built around the principle of Unified Memory.
#3 is that we should take the commitment to PCI Express at face value. PCIe 5 (and Thunderbolt 5) is a big step forward, and the PCIe 6 specification is also now complete. Those designs are sitting on engineering drawing boards right now. An M3 Ultra Mac Pro with PCIe 5 will be a big step beyond the M2 edition. But that would still be the entry-level Mac Pro. We'll just have to wait and see if and when they go beyond that. It really depends on what they do with the M3 Pro/Max graphics.
With the introduction of the Apple Silicon model, some elements of the Mac Pro's utility became less impressive, in part due to the existence of the Mac Studio. With both the Mac Pro and Mac Studio able to use Apple Silicon, and with the Mac Studio being generally cheaper to purchase, it made the Mac Pro a hard purchase for anyone just wanting high levels of performance.
If Apple were to continue making the Mac Pro, it would need to address the major shortcomings that were key to the model's existence for it to become a success. For the moment, if Gurman's comment turns out to be a true indicator, Apple may step back from the Mac Pro for a long time to regroup, or potentially give up on the Mac Pro entirely in its current form.
Apple has talked about the decisions behind the Studio and Pro in interviews. The Mac Studio will be the mainstream, high volume Pro Mac going forward, effectively a continuation of the 2013 Mac Pro that is the best form factor for most professional work and even more affordable than the 2012 model starting at $2k.
The Mac Pro tower is just for professional workflows that rely on PCIe connectivity, mainly high bandwidth video/audio/network IO and for large, fast internal storage. If those use cases didn't exist, they could EOL it.
Both Mac Studio and Pro will probably receive updates when the Ultra chips are due. Next Ultra update isn't likely before October 2024 (14 months out).
The Mac Pro has been a MASSIVELY overpriced after-thought that they toss out a new model just often enough to keep the hardest of the hard-core on the hook since the last of the cheese graters was released in 2012 - and even that was 2-3 years since the previous update, iirc. It's hard not to imagine the designers even surprising themselves at how much they can get people to pay for these machines, and even more so, upgrades and accessories. Even Monster Cable would blush at some of this stuff. ($750 for a set of coaster wheels... I'd REALLY like to see a cogent justification for that. Not that the wheels aren't that good - maybe they are - but why you'd need such expensive precision quality to wheel around a computer!)
Most low shipment volume products are priced much higher than normal. The Mac Pro is already < 100k/year unit volume and the wheels will be for less than half of those. RED does the same for camera equipment:
The Mac Pro isn't overpriced in the sense that Apple makes a massive profit from it. The unit volume is so low that it's not worth making unless they price it like this. They did the same with the 17" MBP.
The Mac Pro isn’t in danger due to a fumbled launch. It’s in danger due to a lack of investment and lack of vision.
In the real world, it’s the people who fumbled the launch who are in danger. Not the product. That usually gets put into more capable hands and the cause of the fumble (humans) gets dealt with.
But I don’t think this was a fumble. Apple purposefully limited it. They could have delayed it and made it great if they wanted it to be great. Instead it’s only Mac Studio level. Sure the studio is great, but it’s performance isnt blowing the doors off high end PCs the way the Mac Pro used to do.
Hopefully it’s just a half measure until the m extreme chips are viable - or multiple SOCs can connect via a system level fabric. But unless the current Mac Pro is upgradeable to multiple SOCs or an Extreme setup, the customer trust will be so badly broken that Apple will have to make it their Zune. The trash can was bad enough. Then they never really kept the the promise of 2019 Mac Pro upgrades outside of a cpu and GPU update and then nothing. And everyone who spent on modules is out a ton of cash. If the new system is all it’s going to be, then shame on Apple. The high end computing company is missing out at the highest end. The Mac Pro was the halo machine. Now it’s just an embarrassment t. How sad.
The Mac Pro isn’t in danger due to a fumbled launch. It’s in danger due to a lack of investment and lack of vision.
In the real world, it’s the people who fumbled the launch who are in danger. Not the product. That usually gets put into more capable hands and the cause of the fumble (humans) gets dealt with.
But I don’t think this was a fumble. Apple purposefully limited it. They could have delayed it and made it great if they wanted it to be great. Instead it’s only Mac Studio level. Sure the studio is great, but it’s performance isnt blowing the doors off high end PCs the way the Mac Pro used to do.
Hopefully it’s just a half measure until the m extreme chips are viable - or multiple SOCs can connect via a system level fabric. But unless the current Mac Pro is upgradeable to multiple SOCs or an Extreme setup, the customer trust will be so badly broken that Apple will have to make it their Zune. The trash can was bad enough. Then they never really kept the the promise of 2019 Mac Pro upgrades outside of a cpu and GPU update and then nothing. And everyone who spent on modules is out a ton of cash. If the new system is all it’s going to be, then shame on Apple. The high end computing company is missing out at the highest end. The Mac Pro was the halo machine. Now it’s just an embarrassment t. How sad.
The Mac Pro never used to do this, it used the same parts as high-end PCs. The Ultra chips are the same performance as the top-end single GPU 2019 Mac Pro.
The equivalent today would be Xeon W9-3495X (2x M2 Ultra CPU) with 48GB VRAM W7900 (60TFLOPs, in real-world examples this is just above M2 Ultra GPU, M1 Ultra = 3080, M2 Ultra = 4080 = 3090, 7900XTX = W7900 between, 4090 = 2x 3080)
So for the price of $6k base + $7k CPU + $5k GPU = $18k, you get up to 2x faster than the $7k M2 Ultra, not that impressive.
M3 Extreme would only be 2x Ultra GPU. This would rival an Nvidia 4090 but M3 Ultra will be 75% of Nvidia 4090 performance for $7k in 2024.
There's always been this tech obsession with the latest hardware 'blowing the doors off' some other model and it usually amounts to being up to 2x faster but only a tiny fraction of people buy the top-end hardware. Apple doesn't need to cater to this.
Apple makes workstations, which only need to do things in real-time. For special purpose tasks, there are options for that like the following 8x Nvidia 4090 (> 600 TFLOPs), AMD Threadripper 5995WX £40k ($50k):
No matter what Apple does, there will be options that are 8x faster or maybe more.
If Apple made solutions like this, they'd sell a handful of units. They wouldn't gain financially and they wouldn't gain brand recognition because the high-end computing space has zero mainstream appeal.
The Mac Studio and Macbook pro can handle the production and raw computing can be done on the compute box.
There was an article recently with comments from people who have used Mac Pros all the time and switched to Macbook Pros because the new M-series chips are fast enough (M2 Max MBP = mid-range 2019 Mac Pro):
There's less demand for higher-end desktops beyond Max/Ultra because the cheaper models can handle the majority of high-end tasks like 4K video editing, coloring, 2D VFX.
3D VFX, AI, scientific computing requirements can go above this and people should invest in the appropriate hardware, a 2x performance boost doesn't change that buying decision. Apple has said they aren't trying to cater to everything, they are doing what they do best - sleek, fast hardware with good user experience and the performance level they target covers 95-99% of the professional computer industry.
The biggest fail of all with the 2023 Mac Pro is it doesn’t have enough PCIE lanes to support all of its slots. Add one 16x card and you’ll be fine, but add a second and they each get only 8x. [...]
This comment (from page one) got six "Informative" votes, which is indicative of just how little people know about what the difference between PCIe 3 versus PCIe 4 means.
Both pools, a total of 24 lanes, are available to all cards. Expansion Slot Utility will allocate it for you, or you can set it manually.
But, what the above comment doesn't grasp is the fact that even high-end GPU cards don't completely saturate 16 PCIe 3 lanes. With PCIe 4 (which doubles the bandwidth), you only need 8 lanes for those cards. So 24 PCIe 4 lanes are the equivalent of 48 PCIe 3 lanes. Not to mention the fact that the M2 Ultra Mac Pro doesn't even support GPU cards, so the whole question is moot. Nothing available today for those x16 slots in the Mac Pro is going to need more than 8 PCIe 4 lanes. Most will need far less.
There's another aspect to this as well, which pertains to the Mac Studio versus Mac Pro question. Obviously, no single external PCIe enclosure connected via Thunderbolt is going to be able to compete with the Mac Pro internal slots. There's a definite subset of users for whom the $3,000 difference makes sense, in terms of both cost and convenience.
Comments
The silicon would have to go beyond the Ultra. Beyond the beyond (ultra means "beyond"). Apple has said enough to make it clear that they did design such a tier, but it's not clear what that was. All that is known is that it was aborted and its code name might have been Jade 4C-Die.
But there are some things that we can surmise:
#1 is that killing Jade 4C-Die wasn't a financial decision, it was about performance. It didn't give them what they were looking for at the high-end Mac Pro starting price of $14K (i.e., double the entry-level price, like the Mac Studio). My assumption is that the reality of the fundamental limits in M1/M2 graphics performance meant its time had not come.
#2 is that Unified Memory is here to stay. Whatever Jade 4C-Die was, and whatever it will become, it will be built around the principle of Unified Memory.
#3 is that we should take the commitment to PCI Express at face value. PCIe 5 (and Thunderbolt 5) is a big step forward, and the PCIe 6 specification is also now complete. Those designs are sitting on engineering drawing boards right now. An M3 Ultra Mac Pro with PCIe 5 will be a big step beyond the M2 edition. But that would still be the entry-level Mac Pro. We'll just have to wait and see if and when they go beyond that. It really depends on what they do with the M3 Pro/Max graphics.
The equivalent today would be Xeon W9-3495X (2x M2 Ultra CPU) with 48GB VRAM W7900 (60TFLOPs, in real-world examples this is just above M2 Ultra GPU, M1 Ultra = 3080, M2 Ultra = 4080 = 3090, 7900XTX = W7900 between, 4090 = 2x 3080)
So for the price of $6k base + $7k CPU + $5k GPU = $18k, you get up to 2x faster than the $7k M2 Ultra, not that impressive.
M3 Extreme would only be 2x Ultra GPU. This would rival an Nvidia 4090 but M3 Ultra will be 75% of Nvidia 4090 performance for $7k in 2024.
There's always been this tech obsession with the latest hardware 'blowing the doors off' some other model and it usually amounts to being up to 2x faster but only a tiny fraction of people buy the top-end hardware. Apple doesn't need to cater to this.
Apple makes workstations, which only need to do things in real-time. For special purpose tasks, there are options for that like the following 8x Nvidia 4090 (> 600 TFLOPs), AMD Threadripper 5995WX £40k ($50k):
https://www.renderboxes.com/product/the-molecule-air-xl/
No matter what Apple does, there will be options that are 8x faster or maybe more.
If Apple made solutions like this, they'd sell a handful of units. They wouldn't gain financially and they wouldn't gain brand recognition because the high-end computing space has zero mainstream appeal.
The Mac Studio and Macbook pro can handle the production and raw computing can be done on the compute box.
There was an article recently with comments from people who have used Mac Pros all the time and switched to Macbook Pros because the new M-series chips are fast enough (M2 Max MBP = mid-range 2019 Mac Pro):
https://www.theverge.com/23770770/apple-mac-pro-m2-ultra-2023-review
There's less demand for higher-end desktops beyond Max/Ultra because the cheaper models can handle the majority of high-end tasks like 4K video editing, coloring, 2D VFX.
3D VFX, AI, scientific computing requirements can go above this and people should invest in the appropriate hardware, a 2x performance boost doesn't change that buying decision. Apple has said they aren't trying to cater to everything, they are doing what they do best - sleek, fast hardware with good user experience and the performance level they target covers 95-99% of the professional computer industry.
https://support.apple.com/en-us/HT213663
Both pools, a total of 24 lanes, are available to all cards. Expansion Slot Utility will allocate it for you, or you can set it manually.
But, what the above comment doesn't grasp is the fact that even high-end GPU cards don't completely saturate 16 PCIe 3 lanes. With PCIe 4 (which doubles the bandwidth), you only need 8 lanes for those cards. So 24 PCIe 4 lanes are the equivalent of 48 PCIe 3 lanes. Not to mention the fact that the M2 Ultra Mac Pro doesn't even support GPU cards, so the whole question is moot. Nothing available today for those x16 slots in the Mac Pro is going to need more than 8 PCIe 4 lanes. Most will need far less.
There's another aspect to this as well, which pertains to the Mac Studio versus Mac Pro question. Obviously, no single external PCIe enclosure connected via Thunderbolt is going to be able to compete with the Mac Pro internal slots. There's a definite subset of users for whom the $3,000 difference makes sense, in terms of both cost and convenience.