Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
if they could produce such a machine they would charge however much it costs to make a profit

look at the price of a studio vs pro

the Mac Pro case itself can't possibly cost them 4 to 5 thousand dollars. profit margin is not an issue

my take is that they have maxed out what they can do with their laptop soc
Sure, but the million dollars question is that would there be enough sales?

Edit: I beg to differ on them already maxing out their SoC capabilities. I think they are just getting started. Guess we’ll just have to wait and see.

Also, computing is getting increasing mobile. Getting your solution prepared for the mobile future is the way to go, I.e. skating to where the puck will be going to. Current solutions from Intel and AMD is, IMHO not suitable.
 
  • Like
Reactions: Pinkyyy 💜🍎
Desktop PC is not dying. Not even close.
Yeah it's not dying. Statistics vary. Shipments may be going down, but revenue appears flat, suggesting a shift to more expensive desktops. "Dying" is hyperbole.

1691809770598.png


1691809869283.png
 
Last edited:
Sure, but the million dollars question is that would there be enough sales?

Edit: I beg to differ on them already maxing out their SoC capabilities. I think they are just getting started. Guess we’ll just have to wait and see.

Also, computing is getting increasing mobile. Getting your solution prepared for the mobile future is the way to go, I.e. skating to where the puck will be going to. Current solutions from Intel and AMD is, IMHO not suitable.

there would be just as many sales as the current Mac Pro, if not more


I'm curious, in your opinion, what is the point of the current Mac Pro?

if they haven't maxed out their soc, why did they release a tower with a laptop board inside
 
Yeah it's not dying. Statistics vary. Shipments may be going down, but revenue appears flat, suggesting a shift to more expensive desktops. "Dying" is hyperbole.

View attachment 2244772

View attachment 2244774

The links he provided earlier are obviously due to covid lockdown, Intel being slow to react to competition, GPU scalpers, and Nvidia's BS 40 series. It's been a crazy several years.
 
  • Like
Reactions: Pinkyyy 💜🍎
there would be just as many sales as the current Mac Pro, if not more


I'm curious, in your opinion, what is the point of the current Mac Pro?

if they haven't maxed out their soc, why did they release a tower with a laptop board inside
Apple would know their target audience. I would not be foolish enough to claim to know better than Apple on their customer base.

Apple has always provided solutions that one up their previous offerings. The current Mac Pro did that. I don’t remember Apple always providing the best of the best in terms of performance. That would be a foolish goal. Expecting Apple to crush their competition is not realistic IMHO. Apple achieving their target performance bracket would be the business decision to be made. If that goal happens to outclass the competition, that’s a bonus that they can brag about. Otherwise they just sell base on their strength, and strength is not just based on brute force performance.

I will have to admit that I am not the target audience of the Mac Pro because my M1 Mini is more than enough for what I need.
 
  • Like
Reactions: Pinkyyy 💜🍎
Apple has always provided solutions that one up their previous offerings. The current Mac Pro did that.

the current Mac Pro in no way "one ups" their previous offering.

despite the quite modest performance boost considering the 4 year interval, it's a downgrade in pretty much every other way
 
Its the smallest market in computers, and rapidly shrinks with each generation of GPUs. The only moments when there are spikes in volume being bought were and maybe still will be crypto booms.

Desktop DIY is dying. And it will die even more when very soon, Small Form Factor PCs, that use AMD APUs or Intel SOCs will become good enough for majority of people that it will be simply not feasible to built your own computer. DIY will become a nieche in current nieche.
We’re looking at 10years minimum. For anything like that to happen.
Those APUs are amazing, the raw power they are spitting out. When I first used a machine that had an integrated gpu. I was NOT impressed.
But now I have a steam deck and it playing Cyberpunk at descent setting. Just a few years ago I would have laughed you out the door. But you are right, the segment is shrinking.
But it won’t go away for a very long time. Professionals and enthusiasts want to be able to build their devices. Have a personal connection with them. Be able to swap out parts and upgrade as needed. With the advent of gpu modules we’re seeing for small form factor machines. It’s gonna be a fun future as far as this is concerned.
 
  • Like
Reactions: Pinkyyy 💜🍎
the current Mac Pro in no way "one ups" their previous offering.

despite the quite modest performance boost considering the 4 year interval, it's a downgrade in pretty much every other way
Well, everyone is entitled to their opinions.
 
DIY SFF, a niche within a niche; make it watercooled, add another niche; make it a Hackintosh, another niche...

Nicheception...
 
Probably not.

Edit:
Maybe just a little.


I like where Apple is heading. Balanced engineering overall. No skewing to any one direction.

Like I said earlier, Apple is just getting started.

Yes it’s faster than its four year old predecessor.

I already conceded that

You also cannot not add any gpu, you cannot add any ram,
And it offers no performance advantage over the Mac Studio

I suppose it’s a matter of opinion but that doesn’t seem like “one upping”
 
Last edited:
  • Like
Reactions: Pinkyyy 💜🍎
And it offers no performance advantage over the Mac Studio
Changing goal post? Wasn't the point of discussion is over whether the current AS Mac Pro has one upped the previous gen Intel Mac Pro?

You also cannot not add any gpu, you cannot add any ram,
And maybe Apple understand their customers that this is not such a requirement?

I would think that those who wanted this is never going to get a Mac Pro. I would think anyone getting a Mac Pro (whether Intel or AS) will be using it for doing profitable work now and depreciating it over 3-5 years. By that time, the next version of the Mac Pro will be way better.
 
Can you provide any reliable, independent technical source that supports your claim?
Have you seen even ONE disassembly of any M1 and M2 series Apple computer?

If you didn't I suggest looking at them and then making your own conclusions.
 
Changing goal post? Wasn't the point of discussion is over whether the current AS Mac Pro has one upped the previous gen Intel Mac Pro?

Indeed that is the point,

the intel Mac Pro was much more powerful than any other Mac at the time of its release, the current as Mac Pro is not

That’s not changing goal posts. That’s a direct comparison between the two
 
  • Like
Reactions: Pinkyyy 💜🍎
you seem very confident of your repetition of this point

what are you basing your opinion on exactly?
Based on: direction of sales of desktop computers, based on direction of sales of PC parts in the market, based on growing sales of mobile computers, based on direction of Intel and AMD development, and based on direction of technical limits of smaller nodes, in which largest die possible may very soon be 400 mm2 from current 900 mm2, based on piling manufacturing and design costs of chips, while also booming game console market and handheld market, while at the same time - upcoming paradigm shift in how we interact with ocmputers, which Apple Vision Pro is only pointing us. Voice control based interaction with computers, gesture control, AI intergation on OS level. That is coming. VERY rapidly.

Intel invested in dGPU to have drivers ready so they can provide integrated solution, like Apple does. AMD - the same thing, for yuears but only with Strix Point in 2024 we will see finally new design with wider memory bus, and pretty large GPU integrated into the SOC package. But both are investing in AI specifcally to have solutions for the paradigm shift that is coming to how we interact with computers, and you NEED integrated hardware for such solutions, because of the AI capabilities.

Overall market will evolve, it will have to. I expect that in very near future, Nvidia GPUs will become at best current midrange offerings, but not because they will forget about the gaming market. It will be required because of reticle limits of process nodes.

Thats why(evolution of the market) Nvidia also wanted to buy ARM, to be able to integrate their GPUs into CPUs, because currently their designs are incompartible, which recent AI supercomputer, as they call it, shown, where CPU and dGPU parts are on separate substrats, but connected through fabric. With market being pushed into integrated solutions: Nvidia will have no place on consumer market, which will be solely: Intel + AMD integrated designs, unless for Nvidia, something miracoulus happens.

DIY is dying, on consumer side and is becoming a nieche, enthusiast nieche. There is more and more mobile chips being manufactured, mobile chips - soldered into BGA packages.

So what is the direction here? Using those, same BGA packages, and soldering into typical DIY motherboards. You get all of the benefits of upgradebility with all of the drawbacks of integrated systems. You buy whole platform. Some of them will come with soldered RAM, some of them will come in the form of NUC, SFF computer, some of them will come in the form of AIO, some of them will come in the form of laptop. Doesn't this remind us of what Apple is doing? Why would Intel and AMD do something like this? Because of manufacturing and design costs. You design one solution to scale up from mobile to desktop thermal envelopes. You simplify manufacturing of already extremely complex products, and packages. And don't think that it will not happen. We ALREADY have motherboards with soldered into them mobile chips, like here: https://www.msi.com/Motherboard/PRO-HM570TI-B-I526

Integration of SOCs is the direction of the whole, overall market. Death of desktop DIY is just emanation of changing preferences of customers. Im not saying that Desktop DIY will die completely, never said it exactly(even when I said that it "will die even more"), just that it will become a nieche within a nieche. Only the highest end, enthusiast level with extremely high profit margin, that is supposed to counteract the lower volume of sales, and extremely high manufacturing and design costs.
 
  • Disagree
Reactions: Romain_H
What you think people want has no bearing on the discussion. The point is that these capabilities were removed. Simply by definition this makes it less capable than its predecessor
It’s weird that you define capabilities as being able to say upgrade memory, while most folks who buys Mac Pro will define it as how well it solves the tasks they bought the Mac Pro for.
 
  • Like
Reactions: eldho
It’s weird that you define capabilities as being able to say upgrade memory, while most folks who buys Mac Pro will define it as how well it solves the tasks they bought the Mac Pro for.

I would indeed define both of those as capabilities

I don’t see what’s weird about that
 
  • Like
Reactions: Pinkyyy 💜🍎
Have you seen even ONE disassembly of any M1 and M2 series Apple computer?

If you didn't I suggest looking at them and then making your own conclusions.
Nice try. You post a couple of very poorly-written (to the point of being incomprehensible) technical claims, don't support them with any links to reliable sources, and then, when challeged, try to distract from your inability to provide a clear explanation by saying I should look at teardowns.

If you actually want to give a serious technical argument, here's how you need to approach this:

First, you could start by citing, say Anandtech, which tells us this ( https://www.anandtech.com/show/17024/apple-m1-max-performance-review ) for the M1 series, and is unchanged on the M2:

M1/2 Pro: 2 x 128 bit LPDDR5 memory controllers = 256 bits
M1/2 Max: 4 x 128 bit LPDDR5 memory controllers = 512 bits
M1/2 Ultra: 8 x 128 bit LPDDR5 memory controllers = 1024 bits

Then you'd need to cite another source showing what the bit width per memory channel is on Apple Silicon. That would give you the number of memory channels. Then you'd need to find another source giving the maximum device size per channel for LPDDR5. You multiply that by the number of channels, and you get the max RAM. To help you out, I found this; but you'd need to do your own research to determine if it's applicable to Apple Silicon: https://www.synopsys.com/designware-ip/technical-bulletin/key-features-about-lpddr5.html

That's how it's done. That is, if you want to make a serious technical claim.

But that only gets you to the M1/M2. The problem is that you're claiming you have "proofs" of the max memory for the M3. Given that we have no confirmed information on the M3, the idea that you're claiming "proof" is just bizarre to me:
But they will increase the bus width with M3 series by 50%, which we already have proofs in the form of 36 GB M3 Pro, and 48 GB M3 Max chips.
Also, even if we knew the max RAM on the M3 will be higher, that still doesn't tell us the bus width will increase. For instance, based on your argument, you'd conclude when they went from a max RAM of 64 GB on the M1 Max to 96 GB on the M2 Max, they must have increased the bus width by 50%. But that's not what happened. The bus width stayed the same. They merely increased the RAM size per module.
 
Last edited:
  • Like
Reactions: Pinkyyy 💜🍎
I would think that those who wanted this is never going to get a Mac Pro.
If this were true, this very discussion would not exist. Phil Schiller would not have apologized for the Trashcan.

I, for examole, would want this (and lots of folk I know would). Apple literally forced me off their platform
 
  • Like
Reactions: Pinkyyy 💜🍎
Nice try. So you can't cite even a single technical source to back up your claims, and try to hide behind that asking me if I've seen teardowns.

How about instead just giving me a straight answer? Again, can you cite a single reliable independent technical source that backs up your claim about the relationship between RAM size and bus width?

And yes, I have seen teardowns. I've in fact posted many times about iFixIt teardowns on this site. It seems very strange that you think you can tell about the relationship between RAM size and bus width from a teardown.
Because you need to have have signal paths on the motherboard. Apple package does not allow you to do so.

You have cited a source saying that you are not limited to VRAM size by bus width. In the context of GPUs - yes, that is correct. The signal paths on the PCB are double sided, so you can have two memory chips connected to the same signal paths on the GPU PCB - doubling the memory. The same way Xbox Series has 16 and 10 GB. 16 GB on a 320 bit bus, and 10 GB on a 128 bit bus.

In the Microsoft consoles case - the addidtional memory chips are on the other side of the motherboard, connected to one of 5 signal paths. So you have sort of partitioned memory subsystem: one side gives you 10 GB of RAM, the other - 6 GB of RAM. 10 GB is the graphics RAM, 6 GB - system RAM.

But because of how RAM is integrated into package to the Apple SOC - you can't. You would have to redesign whole package. Including the manufacturing of it. And then you would have to redesign also the designs. In case of Apple - you ARE limited to memory bus width.

The source on this topic can be any Electrical engineering textbook on the market.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.