No “Extreme” chip coming to Mac Pro?

Well, this begins to make more sense. If there is no Studio that you have to differentiate with, Apple can get away with no Extreme chip. I would expect the Mac Pro prices to be lower than they were with Intel chips, in that case.



Of course, the source is Gurman who may just be trying to defend the “no extreme chip” rumor, so who knows.
 
Well, this begins to make more sense. If there is no Studio that you have to differentiate with, Apple can get away with no Extreme chip. I would expect the Mac Pro prices to be lower than they were with Intel chips, in that case.



Of course, the source is Gurman who may just be trying to defend the “no extreme chip” rumor, so who knows.
I guess that works but Apple would get a ton of deserved shit for that. And from the article it really does read as just Gurman trying to rationalize his various predictions into something cohesive rather than actually having “inside information” that it’s the studio that’s dead. So … possible … but no particular reason to believe it.
 
Hmm, and the amount of cooling in the studio sure looked like later generation, hotter chips were planned.

Apple is free to change its plans at any time of course, but yeah - Gurman :rolleyes:

edit: "The Mac Studio has faced some criticism for its cooling system"
from the 9 to 5 article… what have i missed? Unless they refer to the whine that upset some studio owners
 
Last edited:
Well, this begins to make more sense. If there is no Studio that you have to differentiate with, Apple can get away with no Extreme chip. I would expect the Mac Pro prices to be lower than they were with Intel chips, in that case.



Of course, the source is Gurman who may just be trying to defend the “no extreme chip” rumor, so who knows.
This does leave the desktops in a funny position I feel. No idea of the pricing of the new Mac Pro, but I would imagine it won't start lower than the current one. That means that the highest spec desktop for most people is the Mini with the M2 Pro. That feels wrong. No desktop with a Max chip...unless the Mac Pro has one. I can't imagine that so I"m just gonna say that Gurman is wrong!
 
This does leave the desktops in a funny position I feel. No idea of the pricing of the new Mac Pro, but I would imagine it won't start lower than the current one. That means that the highest spec desktop for most people is the Mini with the M2 Pro. That feels wrong. No desktop with a Max chip...unless the Mac Pro has one. I can't imagine that so I"m just gonna say that Gurman is wrong!
Pretty sure it will be cheaper than the old model.
 
Hmm, and the amount of cooling in the studio sure looked like later generation, hotter chips were planned.

Apple is free to change its plans at any time of course, but yeah - Gurman :rolleyes:

edit: "The Mac Studio has faced some criticism for its cooling system"
from the 9 to 5 article… what have i missed? Unless they refer to the whine that upset some studio owners

I think that if they aren’t doing an M2 Studio it won’t be because of cooling. They may just skip back and forth - M2 Pro, M3 Studio, M4 Pro. Or it could be a one-time hiccup caused by the lack of an M2 Extreme; in that case, M2 Ultra in a Studio vs M2 Ultra in a Pro would be awkward, given that there might not be much benefit of the Pro for the vast majority of users since it’s slots won’t be used for GPUs. Even taking into account that it’s nice to have internal space for drives and a card or two (vs TB4 on a Studio), a lot of people who would consider the Pro would just go with the Studio.

If there is going to be an M2 Ultra Studio, I still maintain that the Mac Pro needs an Extreme, a swappable CPU board, multiple CPU sockets, or some other killer feature to justify it versus Studio.
 
My guess, supported with zero data, is Apple will offer both.

If you need slots and a more robust power supply (doesn't need to be 1.4 kw), get the MacPro. If you don't, get the Studio Ultra. Assuming Apple would re-use the current MacPro chassis (regular and rack mount), that would minimize NRE.
 
1675641117668.png


1675641389536.png


1675641089325.png


If gurman “isn’t familiar” with Miguel, that’s sort of a self-own in itself. https://en.m.wikipedia.org/wiki/Miguel_de_Icaza
 
The 2019 Mac Pro. NO leaker expected that machine to be like that.

No one in Apple leak world knows abouts the Mac Pro. it's Apples most secured product, mainly because it's very low volume.

The 2013 Mac Pro was the same too, a total surprise.

Bottom line, don't trust ANY Mac Pro leaks because unlike Apples other products. This Mac Pro WILL surprise people.
 
The 2019 Mac Pro. NO leaker expected that machine to be like that.

No one in Apple leak world knows abouts the Mac Pro. it's Apples most secured product, mainly because it's very low volume.

The 2013 Mac Pro was the same too, a total surprise.

Bottom line, don't trust ANY Mac Pro leaks because unlike Apples other products. This Mac Pro WILL surprise people.
I thought the reason there were no accurate leaks about the 2019 Mac Pro (at least I don't recall any) is that it was the only product assembled in Apple's own facility in Austin, TX. All others (at the time) were assembled in China. [Yes, many of the Mac Pro's parts were manufactured by foreign contract plants.] That reportedly won't be the case for the Apple Silicon Mac Pro, which will be assembled in Vietnam. When it comes out, and we can assess the accuracy of leaks about it, we'll be able to see if that change made a difference.


Having said that, Apple is getting better at controlling leakers generally, even for products assembled by foreign contract plants. E.g., I don't think anyone knew about the notch on the MBP.
 
A lot what leakers (and generally futurists) do is jus try to guess the most plausible scenario given very limited information. If it is true that there won't be an M2 Extreme (which I personally won't find surprising at all) AND there will be an M2 Mac Pro (which I am not at all convinced about), it only makes sense for there to be no M2 Studio.

All thing considered, it seems very likely that Apple's roadmap was massively messed up by a variety of factors and what we see right now is them trying to salvage the scraps they have. Looking back, I believe M1 and M2 families were intended as initial experiments (M1 being a straightforward application of the existing iPad technology and M2 bringing more desktop-oriented features), and that the "real" Mac hardware was planned for fall 2022 with TSMC's 3nm. That's also why Apple spoke about a "two year transition". M2 family was supposed to be the 2021 chip. But now with TSMC suffering delays and N3 being extremely expensive, I wouldn't be surprised if Apple execs are frantically reshuffling the roadmap again. Irregular product updates, weird products that feel like a one-time thing, awkward execution. It will likely take years until we are back on track.
 
Looking back, I believe M1 and M2 families were intended as initial experiments (M1 being a straightforward application of the existing iPad technology and M2 bringing more desktop-oriented features), and that the "real" Mac hardware was planned for fall 2022 with TSMC's 3nm.
I think you're starting from a somewhat wrong premise here.

Remember, the Mac mini developer transition kits used A12Z. There's at least two Mac-only features Apple hid in these chips: TSO mode (in the PCPUs only), and support for 16GB of DRAM. There's probably more, those are just the obvious ones.

Thus, A12 (and possibly earlier generations?) was when Apple was experimenting. They didn't put in everything needed to launch a real Mac, especially when it comes to I/O, but A12 had enough to be useful as a proof of concept and software development platform.

The M1 family was designed as a "real" Mac family. Some features in it aren't just Mac only, they're desktop Mac only - I'm thinking of the Ultra Fusion bridge in H13J (the codename for M1 Pro/Max/Ultra). Another example: M1 GPU cores aren't identical to A14 GPU cores. I don't remember all the details of that, but one significant change was twice the FP32 compute throughput.
 
I think you're starting from a somewhat wrong premise here.

Remember, the Mac mini developer transition kits used A12Z. There's at least two Mac-only features Apple hid in these chips: TSO mode (in the PCPUs only), and support for 16GB of DRAM. There's probably more, those are just the obvious ones.

Thus, A12 (and possibly earlier generations?) was when Apple was experimenting. They didn't put in everything needed to launch a real Mac, especially when it comes to I/O, but A12 had enough to be useful as a proof of concept and software development platform.

The M1 family was designed as a "real" Mac family. Some features in it aren't just Mac only, they're desktop Mac only - I'm thinking of the Ultra Fusion bridge in H13J (the codename for M1 Pro/Max/Ultra). Another example: M1 GPU cores aren't identical to A14 GPU cores. I don't remember all the details of that, but one significant change was twice the FP32 compute throughput.

Oh, sure, I should have explained myself better. What I was trying to say is that M1 and M2 families are straightforward applications of Apples mobile technology. The base chips even use the packaging Apple has developed for the iPad. You are absolutely right that these chips include multiple features required for the desktop, as well as some desktop-only stuff (UltraFusion) but they all are still horizontally scaled iPhone chips. M1 GPU is still a phone GPU with many limitations (e.g. resister file size) compared to desktop GPUs - btw, I don’t think there is any difference between the A14 and M1, it’s likely A14 throttles issue of FP32 instructions to save power/discourage FP32 use.

I think we will see different, more Mac-centric chip designs going forward, ones that also scale for untethered desktop use. This is where Apple has a clear deficit after all. If they are serious about a Mac Pro, they will have to offer something to offset its likely limited upgradeability. And that something will have to be performance. Looking at the current lineup, an M2 Ultra won’t even compete with enthusiast desktop. Apple needs more vertical scalability. And for that they need cores capable of using more power.
 
Oh, sure, I should have explained myself better. What I was trying to say is that M1 and M2 families are straightforward applications of Apples mobile technology. The base chips even use the packaging Apple has developed for the iPad. You are absolutely right that these chips include multiple features required for the desktop, as well as some desktop-only stuff (UltraFusion) but they all are still horizontally scaled iPhone chips. M1 GPU is still a phone GPU with many limitations (e.g. resister file size) compared to desktop GPUs - btw, I don’t think there is any difference between the A14 and M1, it’s likely A14 throttles issue of FP32 instructions to save power/discourage FP32 use.

I think we will see different, more Mac-centric chip designs going forward, ones that also scale for untethered desktop use. This is where Apple has a clear deficit after all. If they are serious about a Mac Pro, they will have to offer something to offset its likely limited upgradeability. And that something will have to be performance. Looking at the current lineup, an M2 Ultra won’t even compete with enthusiast desktop. Apple needs more vertical scalability. And for that they need cores capable of using more power.
it feels to me that the apple-y solution is to partition GPUs on one chiplet and CPUs on another. That would give them flexibility to scale core counts more appropriately for the specific product, instead of being limited to multiplying everything by powers of 2.
 
it feels to me that the apple-y solution is to partition GPUs on one chiplet and CPUs on another. That would give them flexibility to scale core counts more appropriately for the specific product, instead of being limited to multiplying everything by powers of 2.

Maybe, but they probably still want to keep single-chip solutions for laptops. I can imagine that the die connector comes with a power usage overhead.
 
Back
Top