In three years, a MacBook Air will be faster than this machine.
And the VisionPro will no longer require a headset. Or a battery. You'll wear a beanie with a wind-powered propeller on top, and there will be a holographic heads-up display in the air in front of you. There will be special-edition beanies to support various groups:
Hey, look at that! You are being a raging asshole for no particular reason. At least you are constant.
Instead of jumping to insult someone you might have wanted to look at the performance increase the M2 got and how it compared to the M1 Ultra. If you did that you would have noticed that the comment that you were quick to belittle wasn't that far fetched. In three years the basic M series will mostly certainly out perform the M2 Ultra in single core scores. It will be pretty close on multi.
Less RAM, no real graphics card, and Apple making wild performance claims without any actual benchmarks.
I'm unconvinced. I want to see real benchmarks vs current AMD and Intel processors, and current high end Nvidia and AMD graphics cards. I want to see that because I strongly suspect Apple is just doing a bit of puffery.
The Mac Pro is a kludge. I’m betting they tried to punch a hole in the sky, but the sky punched back: no 3rd party video card support, and no extended RAM to 1.5TB (for serious math/design labs). An M3 Extreme isn’t likely because the niche audience will use PC’s with multiple AMD/nvidia cards or simply off-load to cloud render farms. The days of the Mac Pro are waning - very sad. But the rise of the Mac Studio is a happy compromise. And I agree - stacks of M2 Pro Mac Minis are a serious consideration for homegrown render farms.
Are you able to run or have run some test files that historically use RAM on both an Intel Mac Pro and a new Studio and Pro? It might be interesting to see if the unified memory architecture can actually keep up with an Intel RAM-intensive system. I've seen people keep complaining about this but Apple's Ultra SoC is extremely integrated with a very fast internal bus. It would be interesting to see the results.
I'd also like to see Apple spend a little bit of time working with the TOP500 supercomputer benchmarking software to see how the Ultra compares to GPU-based supercomputers. Apple actually held four spots in the TOP500 list back in 2005 with the cluster system (512 dual-processor G5 Xserves) from Virginia Tech in 14th place with theoretical peak performance estimated at 20240 gigaflops (20.24 Tflops). The M2 Ultra delivers 27.2 teraflops of graphics performance so one current Mac Studio/Pro is faster than the 2005 system.
The new Mac Pro seems to not be what was hinted at from 2 years ago since it’s barely more capable than the Mac Studio but at nearly 3x the cost. It appears Apple failed to deliver on the M2 Extreme processor (2xM2 Ultras) and the M3 pro level chips are most likely nearly a year away from release so Apple was in a bind and had to release a Mac Pro this year and this is what we get. I can’t imagine they will sell many of these machines maybe a few 10’s of thousands. The next M3 iteration of this machine will be the one to buy.
What people keep failing to grasp is that this is the size of the target audience no matter what Apple does. There's no opportunity in this space for some kind of miracle comeback tour at the $5k+ price point.
The biggest workstation manufacturer in the world sells $1-2b of workstations.
If Apple knocked it out the park and sold $1b of these at ASP $7k, they'd sell 140k units vs 20+ million other Macs.
If Apple made a Mac Pro that supported quad Nvidia 4090 GPUs (or equivalent), 2TB RAM, quad M-series CPUs, they'd sell... maybe a few tens of thousands of units, just the same as if they don't offer that option. Likewise with a souped up M3 Extreme.
The fact they dropped MPX modules shows that for all the bloviating about how the 2019 model was the right design for pros, nobody bothered buying them. Meanwhile Apple mentioned they've sold millions of Mac Studios with the exact same design constraints as the 'trashcan'. Phil Schiller's innovative ass will live on.
The Mac Pro is a kludge. I’m betting they tried to punch a hole in the sky, but the sky punched back: no 3rd party video card support, and no extended RAM to 1.5TB (for serious math/design labs). An M3 Extreme isn’t likely because the niche audience will use PC’s with multiple AMD/nvidia cards or simply off-load to cloud render farms. The days of the Mac Pro are waning - very sad. But the rise of the Mac Studio is a happy compromise. And I agree - stacks of M2 Pro Mac Minis are a serious consideration for homegrown render farms.
Another riff on a frequent theme in Apple commentary of late: this inane notion that Apple is pumping out "stopgap" products because this or that didn't quite work out so they had to rush "something" to market. Everything from the Apple Watch 7 (a last minute kludge because the constantly rumored flat-sided design failed in production) to the Mac Studio (a stopgap because the Mac Pro wasn't done in time) to the new Mac Pro itself. Anyone who knows anything about the time needed for R&D, production design and testing, supply chain procurement, etc to bring a tech product to market that meets Apple standards of near perfection knows that these "stopgap product" claims are ridiculous.
The simple storyline usually goes like this: something is rumored (the M2 Exteme! the M3 Hyperdrive!). Said rumored thing doesn't appear in a product as rumored. Therefore, Apple couldn't get said rumored thing to work in time or work at all, so it banged out a stopgap or kludge product last minute. It's just stupid.
Apple's high end computers now seem logically divided. If you need power but not expansion, the base Studio M2 Ultra is $4K. If you need power plus more ports and card expansion capability, the base Mac Pro is $7K. In terms of upspec'ing from the base model, the costs are the same for the Studio and Pro. Maxed out, your $12K investment in a Mac Pro now buys you a machine that will easily outperform what cost $54.000 previously... and people are disappointed with this. It just never ends.
The Mac Pro is a kludge. I’m betting they tried to punch a hole in the sky, but the sky punched back: no 3rd party video card support, and no extended RAM to 1.5TB (for serious math/design labs). An M3 Extreme isn’t likely because the niche audience will use PC’s with multiple AMD/nvidia cards or simply off-load to cloud render farms. The days of the Mac Pro are waning - very sad. But the rise of the Mac Studio is a happy compromise. And I agree - stacks of M2 Pro Mac Minis are a serious consideration for homegrown render farms.
Another riff on a frequent theme in Apple commentary of late: this inane notion that Apple is pumping out "stopgap" products because this or that didn't quite work out so they had to rush "something" to market. Everything from the Apple Watch 7 (a last minute kludge because the constantly rumored flat-sided design failed in production) to the Mac Studio (a stopgap because the Mac Pro wasn't done in time) to the new Mac Pro itself. Anyone who knows anything about the time needed for R&D, production design and testing, supply chain procurement, etc to bring a tech product to market that meets Apple standards of near perfection knows that these "stopgap product" claims are ridiculous.
The simple storyline usually goes like this: something is rumored (the M2 Exteme! the M3 Hyperdrive!). Said rumored thing doesn't appear in a product as rumored. Therefore, Apple couldn't get said rumored thing to work in time or work at all, so it banged out a stopgap or kludge product last minute. It's just stupid.
Apple's high end computers now seem logically divided. If you need power but not expansion, the base Studio M2 Ultra is $4K. If you need power plus more ports and card expansion capability, the base Mac Pro is $7K. In terms of upspec'ing from the base model, the costs are the same for the Studio and Pro. Maxed out, your $12K investment in a Mac Pro now buys you a machine that will easily outperform what cost $54.000 previously... and people are disappointed with this. It just never ends.
I could not agree more that it makes little sense to hold Apple to expectations that are are based on rumors and predictions that are generated by people outside of Apple. In many cases the prognosticators attach cute names to products and components that have no basis in Apple’s naming scheme. There are probably at least a half dozen predicted names for what Apple revealed as the Vision Pro. Should we be planning a funeral service for all those now reliably dead names?
The only real expectation that came directly from Apple was that they would move their whole Mac product line over to Apple Silicon. It finally happened with the release of the Apple Silicon Mac Pro. is it a drop-in replacement for the Intel Mac Pro? Nope, just like the 24” iMac is not a drop-in replacement for the 27” iMac. Apple gets to decide where it wants to focus its investments and how it wants to structure its portfolio going forward. It is what it is and they seem to be content with what it is, at least at this point in time. They fully reserve their right to change their mind, add new stuff, delete old stuff, or make tweaks that emphasize the attractiveness their most profitable and/or populate products, like the MacBook Air.
Apple may very well have decided to concede a segment of the market, e.g., the very highest of the high-end workstations, to further refine their focus in other areas. This is no different than what Ford Motor Company did when dropping all but one of it cars (Mustang) from its product portfolio and focusing almost exclusively on SUVs and trucks. Those who had their heart set on getting an updated Fiesta, Focus, Escort, Taurus, are SOL, just like many police forces (and old men) were very disappointed when Ford dropped the Crown Victoria. Not only can companies drop products, they can also redefine products, as anyone who’s bought a Ford Mustang Mach-E clearly knows.
On Mac Studio M2 Ultra, I see 4 TB ports on rear, 2 TB ports on front, and 1 HDMI port on rear. So no way to directly connect "Eight displays with up to 4K resolution at 60Hz" ...at least one TB dual-display adapter would be needed.
If I want to drive 8 HDMI displays, it should be possible to install four TB-to-dual HDMI adapters on the 4 rear M2 Ultra TB ports.
I'll be looking into this further, but if anyone has any insights on doing this, I'd appreciate hearing about it.
On Mac Studio M2 Ultra, I see 4 TB ports on rear, 2 TB ports on front, and 1 HDMI port on rear. So no way to directly connect "Eight displays with up to 4K resolution at 60Hz" ...at least one TB dual-display adapter would be needed.
If I want to drive 8 HDMI displays, it should be possible to install four TB-to-dual HDMI adapters on the 4 rear M2 Ultra TB ports.
I'll be looking into this further, but if anyone has any insights on doing this, I'd appreciate hearing about it.
I think the key is daisychaining. Google it.
I know what daisy chaining is. You’re assuming I am using TB monitors. I’m not. My applications use commercial monitors for video wall type arrays or multiple projectors for projection mapping.
(Many commercial displays can be daisy chained, but this is for distributing the same signal to multi-monitor arrays. If you don’t need to address the native resolution of the entire array, this can work.)
PS: I appreciate you taking the time to post, but if you’re going to make a suggestion, a brief explanation or link would be helpful. Just saying “Google it” is lazy.
Less RAM, no real graphics card, and Apple making wild performance claims without any actual benchmarks.
I'm unconvinced. I want to see real benchmarks vs current AMD and Intel processors, and current high end Nvidia and AMD graphics cards. I want to see that because I strongly suspect Apple is just doing a bit of puffery.
I strongly suspect you’re not in the market for this machine and thus are just complaining yet again. If only we had a word for such a thing…
Mac Pro 2023 is all about PCIe. Apple decided to stay on PCIe 4 while Dell, HP,... are shipping workstations with double bandwidth PCIe 5.0. The real question for heavy Pro users is if Apple Silicon is good for them.
Buy Mac Pro 2023 and get half speed PCIe.
Wait 1 year for M3 Ultra as it might deliver PCIe 5. However, the main reason for Apple to support PCIe 5 is Thunderbolt 5, and that standard won't be ready in time for M3; PCIe 5.0 might not be part of M3.
Replace old Xeon-based workstation (from Apple) with new Xeon running other OS, get the latest PCIe, and keep the options to choose GPUs and expand memory as required.
The direction of Apple Silicon is clear. Pro = Studio in a PCIe 4 enclosure with no other modularity. The Pro crowd will have to make a decision.
The new Mac Pro seems to not be what was hinted at from 2 years ago since it’s barely more capable than the Mac Studio but at nearly 3x the cost. It appears Apple failed to deliver on the M2 Extreme processor (2xM2 Ultras) and the M3 pro level chips are most likely nearly a year away from release so Apple was in a bind and had to release a Mac Pro this year and this is what we get. I can’t imagine they will sell many of these machines maybe a few 10’s of thousands. The next M3 iteration of this machine will be the one to buy.
1. Let us know how many expansion cards you can shove inside that Mac Studio, will you?
2. How — exactly — can Apple “fail to deliver” something they never, ever promised, Mr. Entitled? Did you know that your theoretical fantasy chip involves exponentially more work to create than just hot-gluing two M2 Ultra’s together?
3. You may be right about the sales predictions, but my own theoretical fantasy regarding this machine is that now that its out, Apple might strike a deal with Nvidia — or heck, just make in-house — video cards that can with, rather than be a substitute for, the incredible on-board GPU. I think those PCI slots will eventually prove very useful in unexpected ways, though again you may be right that that’s a job for the M3 or later.
In three years, a MacBook Air will be faster than this machine.
And the VisionPro will no longer require a headset. Or a battery. You'll wear a beanie with a wind-powered propeller on top, and there will be a holographic heads-up display in the air in front of you. There will be special-edition beanies to support various groups:
I’d rather be where Apple is performance low wattage, with a new ecosystem.
Less RAM, no real graphics card, and Apple making wild performance claims without any actual benchmarks.
I'm unconvinced. I want to see real benchmarks vs current AMD and Intel processors, and current high end Nvidia and AMD graphics cards. I want to see that because I strongly suspect Apple is just doing a bit of puffery.
In all my years of Apple watching and attending events, Apple has never made a 'wild' claim on tech specs. They will pick language carefully - as we all should - what exactly they say, but I've never seen them BS. Not once when it comes to tech ability. And I attended my first MacWorld thingy, or whatever it was called in 1994. And have buying Macs since around 1988.
As a tech company, they may spin and market, but I've never seen the BS that comes out of other companies when it comes to tech specs.
The Mac Pro is a kludge. I’m betting they tried to punch a hole in the sky, but the sky punched back: no 3rd party video card support, and no extended RAM to 1.5TB (for serious math/design labs). An M3 Extreme isn’t likely because the niche audience will use PC’s with multiple AMD/nvidia cards or simply off-load to cloud render farms. The days of the Mac Pro are waning - very sad. But the rise of the Mac Studio is a happy compromise. And I agree - stacks of M2 Pro Mac Minis are a serious consideration for homegrown render farms.
The problem is, unless they are abandoning that market (3D pros), there will have to be a certain amount of GPU performance local. You can't work in the actual app with unresponsive viewports, or inaccurate previews (compared to a PC), with cloud render-farms. There has to be a level of suitable local GPU performance. (Note: most of the benchmarks miss this.)
That said, there have been some good Unreal Engine demos I've seen (they aren't as good as PCs, but seem quite usable), and I'd guess Maya must be performing OK? So, that will make it possible, assuming other apps optimize. Maybe we'll be OK more fully when the M3 brings some RT hardware.
That still leaves competitiveness. This stuff already works well on a $1500 PC, and I can't imagine that tech will just sit still and wait for Apple to catch up. My main concern (at least at this point) is capability, not matching the PC market. But, when it comes time to buy, I'm going to have to carefully consider putting extra money into a Mac Studio, or if I should just buy a mini and a $1500-2000 PC I can remote control in a window. I see little advantage, right now, to buying anything beyond a base-level Mac for what I do.
Less RAM, no real graphics card, and Apple making wild performance claims without any actual benchmarks.
I'm unconvinced. I want to see real benchmarks vs current AMD and Intel processors, and current high end Nvidia and AMD graphics cards. I want to see that because I strongly suspect Apple is just doing a bit of puffery.
Oh, they totally are. And, more than benchmarks. Look at performance within actual apps you use (or intend to use). For certain multi-core CPU stuff, or things that can use the special video-encoding hardware, etc. these machines will scream. But, for things you'd have a PC with a 4080 or such, it probably won't go so well.
rob53 said: The M2 Ultra delivers 27.2 teraflops of graphics performance so one current Mac Studio/Pro is faster than the 2005 system.
Just be sure to check what it actually does in your workflow. How many teraflops it can do is irrelevant if the viewport on your 3D app is too choppy to use.
Marvin said: If Apple made a Mac Pro that supported quad Nvidia 4090 GPUs (or equivalent), 2TB RAM, quad M-series CPUs, they'd sell... maybe a few tens of thousands of units, just the same as if they don't offer that option. Likewise with a souped up M3 Extreme.
The fact they dropped MPX modules shows that for all the bloviating about how the 2019 model was the right design for pros, nobody bothered buying them. Meanwhile Apple mentioned they've sold millions of Mac Studios with the exact same design constraints as the 'trashcan'.
Yes, this is true as far as it goes. It depends on what markets they want to serve. I think it benefits them overall to support the range, but they'll always make more money down on the consumer end than high-end pros, for the most part.
But, I'm not sure I agree with that second statement. I think they moved to this new Mac Pro to fulfill their transition, and probably couldn't figure out a way to do the MPX thing on the new platform, or just don't plan to do it.
I think for all the bloviating about how the 2019 model was the right design for the pros, they just abandoned that all with their new platform direction. (Unless I'm missing something, they could implement AMD cards and eGPUs and such, they've just decided not to.)
charlesn said: Another riff on a frequent theme in Apple commentary of late: this inane notion that Apple is pumping out "stopgap" products because this or that didn't quite work out so they had to rush "something" to market. Everything from the Apple Watch 7 (a last minute kludge because the constantly rumored flat-sided design failed in production) to the Mac Studio (a stopgap because the Mac Pro wasn't done in time) to the new Mac Pro itself. Anyone who knows anything about the time needed for R&D, production design and testing, supply chain procurement, etc to bring a tech product to market that meets Apple standards of near perfection knows that these "stopgap product" claims are ridiculous.
The simple storyline usually goes like this: something is rumored (the M2 Exteme! the M3 Hyperdrive!). Said rumored thing doesn't appear in a product as rumored. Therefore, Apple couldn't get said rumored thing to work in time or work at all, so it banged out a stopgap or kludge product last minute. It's just stupid.
Apple's high end computers now seem logically divided. If you need power but not expansion, the base Studio M2 Ultra is $4K. If you need power plus more ports and card expansion capability, the base Mac Pro is $7K. In terms of upspec'ing from the base model, the costs are the same for the Studio and Pro. Maxed out, your $12K investment in a Mac Pro now buys you a machine that will easily outperform what cost $54.000 previously... and people are disappointed with this. It just never ends.
I disagree. The new Mac Pro only meets the needs of a fraction of the market the old Mac Pro did (which was already small). While rumors and wishes can get a bit wild, I think a basic presumption that the new Mac Pro would be somewhat competent compared to the previous model is warranted. The new one is faster on some metrics, but not even the the ballpark on others (like GPU performance).
My read, is they simply wanted to complete the transition, so they did. They've done this in past transitions. I'd be a bit concerned if I invested in an Intel Mac Pro, as I wonder how long they'll support it OS-wise.
No, a $12k Mac Pro won't even come close to outperforming a previous Mac Pro with a single higher end GPU option, let alone 4 of them. But, the reason for the complaining, is that for people in 3D who use AMD (or Nvidia) based systems, Apple has no competitive machine any longer. Even if you've got unlimited funds, Apple doesn't have the hardware. And, there is little in-between, either. You're better off buying a $1500 gaming PC, and a $600 mini, I guess.
dewme said: The only real expectation that came directly from Apple was that they would move their whole Mac product line over to Apple Silicon. It finally happened with the release of the Apple Silicon Mac Pro. is it a drop-in replacement for the Intel Mac Pro? Nope, just like the 24” iMac is not a drop-in replacement for the 27” iMac. Apple gets to decide where it wants to focus its investments and how it wants to structure its portfolio going forward. It is what it is and they seem to be content with what it is, at least at this point in time.
My read on it, is that the plan seems to be they'll have something more workable when they hit the M3 or M4. If they make enough GPU advancement (and add RT, etc.), it might not match what is going on in the PC world with AMD/Nvidia, but it might be good enough. We'll see.
Less RAM, no real graphics card, and Apple making wild performance claims without any actual benchmarks.
I'm unconvinced. I want to see real benchmarks vs current AMD and Intel processors, and current high end Nvidia and AMD graphics cards. I want to see that because I strongly suspect Apple is just doing a bit of puffery.
I strongly suspect you’re not in the market for this machine and thus are just complaining yet again. If only we had a word for such a thing…
Same for me... true to some extent. I can't afford a Mac Pro. But, the Mac Pro is an indicator of where the platform is going in terms of potential, so I watch it closely.
I had been expecting I'd buy a Mac Studio once it was introduced. It seemed like the perfect machine I've always wanted. I still might buy one when the M3 version is released. The problem the new Mac Pro highlighted, though, is that I still don't see a clear plan for GPU power. I need GPU power. So, if the M3 doesn't cut it, I may be much better off minimizing my Mac cost (ex: fairly base mini), and then buying some real 3D power in a PC. I can just remote-control the PC for 3D apps and gaming, and the mini should be plenty fast enough for the rest of my Mac stuff.
The direction of Apple Silicon is clear. Pro = Studio in a PCIe 4 enclosure with no other modularity. The Pro crowd will have to make a decision.
I guess this is part of the debate. Is this representative of Apple's new direction? Or, is it simply a stop-gap? I'm really hoping it is the latter, because otherwise it is quite disappointing.
chasm said: 1. Let us know how many expansion cards you can shove inside that Mac Studio, will you?
2. How — exactly — can Apple “fail to deliver” something they never, ever promised, Mr. Entitled? Did you know that your theoretical fantasy chip involves exponentially more work to create than just hot-gluing two M2 Ultra’s together?
3. You may be right about the sales predictions, but my own theoretical fantasy regarding this machine is that now that its out, Apple might strike a deal with Nvidia — or heck, just make in-house — video cards that can with, rather than be a substitute for, the incredible on-board GPU. I think those PCI slots will eventually prove very useful in unexpected ways, though again you may be right that that’s a job for the M3 or later.
Ok, you get a 'like' for that 3rd point (and I so hope you're correct). But, re: 1 - True, but what percent of the 2019 Mac Pro user base needed those card slots (besides GPU)? re: 2 - Fair point... except, I think it is reasonable to make the assumption the new Mac Pro will be better than the old one. If you need GPU-power, the old Mac Pro will wipe the floor with the new one. Heck, my 2018 Mac mini will wipe the floor with the new one. (Apple has no mid-to-high-end anymore in this regard.)
Comments
This is a brutal evisceration of the PCie layout
Hector Martin: "So the weird thing is I don't …" - Treehouse Mastodon
Instead of jumping to insult someone you might have wanted to look at the performance increase the M2 got and how it compared to the M1 Ultra. If you did that you would have noticed that the comment that you were quick to belittle wasn't that far fetched. In three years the basic M series will mostly certainly out perform the M2 Ultra in single core scores. It will be pretty close on multi.
https://youtu.be/lLJtqwNZO84
"Seems egregious for a $7k machine"
I'd also like to see Apple spend a little bit of time working with the TOP500 supercomputer benchmarking software to see how the Ultra compares to GPU-based supercomputers. Apple actually held four spots in the TOP500 list back in 2005 with the cluster system (512 dual-processor G5 Xserves) from Virginia Tech in 14th place with theoretical peak performance estimated at 20240 gigaflops (20.24 Tflops). The M2 Ultra delivers 27.2 teraflops of graphics performance so one current Mac Studio/Pro is faster than the 2005 system.
The biggest workstation manufacturer in the world sells $1-2b of workstations.
If Apple knocked it out the park and sold $1b of these at ASP $7k, they'd sell 140k units vs 20+ million other Macs.
If Apple made a Mac Pro that supported quad Nvidia 4090 GPUs (or equivalent), 2TB RAM, quad M-series CPUs, they'd sell... maybe a few tens of thousands of units, just the same as if they don't offer that option. Likewise with a souped up M3 Extreme.
The fact they dropped MPX modules shows that for all the bloviating about how the 2019 model was the right design for pros, nobody bothered buying them. Meanwhile Apple mentioned they've sold millions of Mac Studios with the exact same design constraints as the 'trashcan'. Phil Schiller's innovative ass will live on.
History has a way of repeating itself:
The simple storyline usually goes like this: something is rumored (the M2 Exteme! the M3 Hyperdrive!). Said rumored thing doesn't appear in a product as rumored. Therefore, Apple couldn't get said rumored thing to work in time or work at all, so it banged out a stopgap or kludge product last minute. It's just stupid.
Apple's high end computers now seem logically divided. If you need power but not expansion, the base Studio M2 Ultra is $4K. If you need power plus more ports and card expansion capability, the base Mac Pro is $7K. In terms of upspec'ing from the base model, the costs are the same for the Studio and Pro. Maxed out, your $12K investment in a Mac Pro now buys you a machine that will easily outperform what cost $54.000 previously... and people are disappointed with this. It just never ends.
The only real expectation that came directly from Apple was that they would move their whole Mac product line over to Apple Silicon. It finally happened with the release of the Apple Silicon Mac Pro. is it a drop-in replacement for the Intel Mac Pro? Nope, just like the 24” iMac is not a drop-in replacement for the 27” iMac. Apple gets to decide where it wants to focus its investments and how it wants to structure its portfolio going forward. It is what it is and they seem to be content with what it is, at least at this point in time. They fully reserve their right to change their mind, add new stuff, delete old stuff, or make tweaks that emphasize the attractiveness their most profitable and/or populate products, like the MacBook Air.
Apple may very well have decided to concede a segment of the market, e.g., the very highest of the high-end workstations, to further refine their focus in other areas. This is no different than what Ford Motor Company did when dropping all but one of it cars (Mustang) from its product portfolio and focusing almost exclusively on SUVs and trucks. Those who had their heart set on getting an updated Fiesta, Focus, Escort, Taurus, are SOL, just like many police forces (and old men) were very disappointed when Ford dropped the Crown Victoria. Not only can companies drop products, they can also redefine products, as anyone who’s bought a Ford Mustang Mach-E clearly knows.
I think the key is daisychaining. Google it.
- Buy Mac Pro 2023 and get half speed PCIe.
- Wait 1 year for M3 Ultra as it might deliver PCIe 5. However, the main reason for Apple to support PCIe 5 is Thunderbolt 5, and that standard won't be ready in time for M3; PCIe 5.0 might not be part of M3.
- Replace old Xeon-based workstation (from Apple) with new Xeon running other OS, get the latest PCIe, and keep the options to choose GPUs and expand memory as required.
The direction of Apple Silicon is clear. Pro = Studio in a PCIe 4 enclosure with no other modularity. The Pro crowd will have to make a decision.3. You may be right about the sales predictions, but my own theoretical fantasy regarding this machine is that now that its out, Apple might strike a deal with Nvidia — or heck, just make in-house — video cards that can with, rather than be a substitute for, the incredible on-board GPU. I think those PCI slots will eventually prove very useful in unexpected ways, though again you may be right that that’s a job for the M3 or later.
As a tech company, they may spin and market, but I've never seen the BS that comes out of other companies when it comes to tech specs.
That said, there have been some good Unreal Engine demos I've seen (they aren't as good as PCs, but seem quite usable), and I'd guess Maya must be performing OK? So, that will make it possible, assuming other apps optimize. Maybe we'll be OK more fully when the M3 brings some RT hardware.
That still leaves competitiveness. This stuff already works well on a $1500 PC, and I can't imagine that tech will just sit still and wait for Apple to catch up. My main concern (at least at this point) is capability, not matching the PC market. But, when it comes time to buy, I'm going to have to carefully consider putting extra money into a Mac Studio, or if I should just buy a mini and a $1500-2000 PC I can remote control in a window. I see little advantage, right now, to buying anything beyond a base-level Mac for what I do.
Oh, they totally are. And, more than benchmarks. Look at performance within actual apps you use (or intend to use). For certain multi-core CPU stuff, or things that can use the special video-encoding hardware, etc. these machines will scream. But, for things you'd have a PC with a 4080 or such, it probably won't go so well.
Yes, this is true as far as it goes. It depends on what markets they want to serve. I think it benefits them overall to support the range, but they'll always make more money down on the consumer end than high-end pros, for the most part.
But, I'm not sure I agree with that second statement. I think they moved to this new Mac Pro to fulfill their transition, and probably couldn't figure out a way to do the MPX thing on the new platform, or just don't plan to do it.
I think for all the bloviating about how the 2019 model was the right design for the pros, they just abandoned that all with their new platform direction. (Unless I'm missing something, they could implement AMD cards and eGPUs and such, they've just decided not to.)
I disagree. The new Mac Pro only meets the needs of a fraction of the market the old Mac Pro did (which was already small). While rumors and wishes can get a bit wild, I think a basic presumption that the new Mac Pro would be somewhat competent compared to the previous model is warranted. The new one is faster on some metrics, but not even the the ballpark on others (like GPU performance).
My read, is they simply wanted to complete the transition, so they did. They've done this in past transitions. I'd be a bit concerned if I invested in an Intel Mac Pro, as I wonder how long they'll support it OS-wise.
No, a $12k Mac Pro won't even come close to outperforming a previous Mac Pro with a single higher end GPU option, let alone 4 of them. But, the reason for the complaining, is that for people in 3D who use AMD (or Nvidia) based systems, Apple has no competitive machine any longer. Even if you've got unlimited funds, Apple doesn't have the hardware. And, there is little in-between, either. You're better off buying a $1500 gaming PC, and a $600 mini, I guess.
My read on it, is that the plan seems to be they'll have something more workable when they hit the M3 or M4. If they make enough GPU advancement (and add RT, etc.), it might not match what is going on in the PC world with AMD/Nvidia, but it might be good enough. We'll see.
Same for me... true to some extent. I can't afford a Mac Pro. But, the Mac Pro is an indicator of where the platform is going in terms of potential, so I watch it closely.
I had been expecting I'd buy a Mac Studio once it was introduced. It seemed like the perfect machine I've always wanted. I still might buy one when the M3 version is released. The problem the new Mac Pro highlighted, though, is that I still don't see a clear plan for GPU power. I need GPU power. So, if the M3 doesn't cut it, I may be much better off minimizing my Mac cost (ex: fairly base mini), and then buying some real 3D power in a PC. I can just remote-control the PC for 3D apps and gaming, and the mini should be plenty fast enough for the rest of my Mac stuff.
I guess this is part of the debate. Is this representative of Apple's new direction? Or, is it simply a stop-gap? I'm really hoping it is the latter, because otherwise it is quite disappointing.
Ok, you get a 'like' for that 3rd point (and I so hope you're correct).
But, re: 1 - True, but what percent of the 2019 Mac Pro user base needed those card slots (besides GPU)?
re: 2 - Fair point... except, I think it is reasonable to make the assumption the new Mac Pro will be better than the old one. If you need GPU-power, the old Mac Pro will wipe the floor with the new one. Heck, my 2018 Mac mini will wipe the floor with the new one. (Apple has no mid-to-high-end anymore in this regard.)