Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
This. This is what changed everything for real pro’s. Apple has abandoned them in software. And so they’ve abandoned Apple and gone to both Windows and Linux.
It wasn't just the pros - it was also the power users.

I was pointing this out when the 1st AS product dropped - the true believers were constantly ranting on how all of these software companies would just drop whatever they were doing and make an AS native product.

Technical illiterates swearing up and down that all it would take was flipping a couple of switches and recompiling the product.
 
I presume Apple could have on-package memory up to a certain limit (<256GB) and then additional DIMM-based memory using an external memory controller (and running at slower speeds with higher latency). So that would allow a significant amount of highest-speed memory for the majority of tasks, but the ability to run truly massive amounts of memory for those tasks that need it (with a performance penalty).
Sure, why not. This kind of thing was done in various computers of the past, like Amigas.
 
The Mini and the Studio each have their own heat limitations. Apple will want to discourage users from forcing usgaes that exceed the inherent heat management capabilities of each different level of desktop Mini-Studio-Pro.
People always seem to refuse to think about thermal throttling when they attack commenters who want pro machines 🤷🏽‍♂️
 
  • Like
Reactions: ssgbryan
[…] If someone wants Intel performance, the best way to get that now and in the future is with Intel systems, likely something custom built by someone that has studied the varied landscape of PC desktop parts and know how to pick precisely which parts mesh well with others to provide that performance.
Ah yes, PC build voodoo.
 
This is why the Mac Pro doesn't make sense to me as a product, at all. If the Mac Pro user is one who is going to peg their CPU for hours or days at a time, then what you're really in need of is raw performance at the expense of everything else. Time is money, and nothing else should matter except how long that CPU needs to be pegged. You should be looking for the absolute fastest way possible that you can afford to grind through that process. And that's almost never going to be a Mac, because as fast as a Mac can be, there are always other priorities for the system that is being packaged and sold other than just raw performance. Always.

The M1 Ultra is fast enough for probably 80% of professional users, and the M2 Ultra sounds like it will be fast enough for 90%+. The remaining class who are all about pegging super computers for content production or whatever it is you're doing...need not to be looking at brands.
Yet again:

THERMALS.
 
  • Like
Reactions: ssgbryan
There is also MacOS. I took over a year to ween myself off it because it is literally all I used at home for years.
Some people just can't not use MacOS or are tied to it for Logic or FCPX still.
I used Windows for most of my life. I will not go back for anything but gaming. I have absolutely zero tolerance for it anymore.
 
People always seem to refuse to think about thermal throttling when they attack commenters who want pro machines 🤷🏽‍♂️
I have a LOT of dead Apple equipment because Sir Idiot Boy didn't understand the concept of heat dissipation.

The other comment I would always get is why do you need that much power? And a side order of Why do you have so much music & movies? You don't need that.
 
Looking forward to the new Mac Pro being released, hopefully with a 3nm Apple Silicon chip.
I'm also hoping that will happen. How I think it could play out is that the M3 chip would come out in October 2023 so it could be based on the 3nm A17 architecture (A15 and A16 based chips wouldn't be 3nm since SoCs of the same architecture wouldn't be manufactured on different nodes). The M3 would come in a new iPad Pro and iMac, while the M3 Ultra and possibly Extreme, the SoCs for the Mac Pro, would be announced at the same time as the regular M3 so that the Apple silicon Mac Pro could launch in December similar to the previous two models. We may or may not get a sneak peek of the Mac Pro at WWDC like we got for the previous two. In that case, the sneak peek would not specify what SoC would be in the Mac Pro.

Never made much sense to only put this in a Mac Pro.
I'm thinking Apple could make M-series Extreme chips for use in the Apple Car. The potential self-driving functionality would be based around machine learning, which a 64-core Neural Engine like what has been rumored for the M2 Extreme would certainly help accelerate. That said, if the self-driving functionality is limited like current rumors suggest, it may not need as powerful of a Neural Engine. Apple could also make automobile-specific chips that have few CPU and GPU cores but massive numbers of Neural Engine cores. But maybe I'm wrong in thinking that more Neural Engine cores is better.

A Mac Studio is not a workstation.
1671718567318.png
 
Last edited:
what the point of Mac Pro with M2 ultra since we already have mac studio for that
Mac Pro it should stand above or way above of Mac Studio
Expandability. PCI slots. Ability to upgrade components such as ram. Ability to upgrade or augment GPU. Not being forced to rely on thunderbolt for peripherals.
 


Apple continues to test an all-new Mac Pro with an M2 Ultra chip, but the company has likely abandoned plans to release a higher-end configuration with a so-called "M2 Extreme" chip, according to Bloomberg's Mark Gurman.


In the latest edition of his newsletter today, Gurman said the Mac Pro with the M2 Ultra chip will be available with up to a 24-core CPU, up to a 76-core GPU, and at least 192GB of RAM. Like the current Mac Pro, he expects the new model to remain expandable, allowing for additional memory, storage, and other components to be inserted.

The higher-end model with the M2 Extreme chip would have been available with up to a 48-core CPU and up to a 152-core GPU, according to Gurman, but he believes that this configuration was scrapped due to cost and manufacturing complexities.

"Based on Apple's current pricing structure, an M2 Extreme version of a Mac Pro would probably cost at least $10,000 — without any other upgrades — making it an extraordinarily niche product that likely isn't worth the development costs, engineering resources and production bandwidth it would require," he wrote.

The current Intel-based Mac Pro was released in December 2019 and starts at $5,999 in the United States. Barring any further delays, the new model will presumably launch at some point in 2023, but Gurman did not provide an updated timeframe.

Article Link: Gurman: All-New Mac Pro Still in Testing, But 'M2 Extreme' Chip Likely Canceled
This ignores the need for a machine in between the small iMac and the Mac Studio
 
I'm thinking Apple could make M-series Extreme chips for use in the Apple Car.
The first Teslas run on Intel Atoms. The newer ones use AMD Ryzen processors (MCU3). A car wouldn’t need M-series Extreme type performance. The only thing that much performance would do is likely just cut the range just like Tesla’s ranges were cut when they went from Atom to Ryzen.
 
The first Teslas run on Intel Atoms. The newer ones use AMD Ryzen processors (MCU3). A car wouldn’t need M-series Extreme type performance. The only thing that much performance would do is likely just cut the range just like Tesla’s ranges were cut when they went from Atom to Ryzen.
I was thinking they could use M-series Extreme chips solely for the sheer number of Neural Engine cores, used to accelerate machine learning for self-driving capabilities, not CPU or GPU cores. So does that ML acceleration just not matter, then? How many Neural Engine cores would actually be necessary for a car with self-driving capabilities limited to places like highways, like the Apple Car is currently rumored to be?
 
I was thinking they could use M-series Extreme chips solely for the sheer number of Neural Engine cores, used to accelerate machine learning for self-driving capabilities, not CPU or GPU cores. So does that ML acceleration just not matter, then? How many Neural Engine cores would actually be necessary for a car with self-driving capabilities limited to places like highways, like the Apple Car is currently rumored to be?
More TOPs for sure. For example the AP computer in Teslas has 36 TOPs per node (there are 2) and by all accounts it isn't enough for self driving. Mercedes is supposed to be using Drive PX Pegasus which can do 320 TOPs. The Mustang MachE using EyeQ4 only has like 2 TOPs.
 
I was thinking they could use M-series Extreme chips solely for the sheer number of Neural Engine cores, used to accelerate machine learning for self-driving capabilities, not CPU or GPU cores. So does that ML acceleration just not matter, then? How many Neural Engine cores would actually be necessary for a car with self-driving capabilities limited to places like highways, like the Apple Car is currently rumored to be?
Excellent point. Does make me wonder why they switched to a more power hungry chip that would actually affect range?

I currently own a ‘non-self-driving” Kia K5 that is effectively self driving for as much as I need it to be on highways. I wonder what level of hardware is currently used on those? I can understand a good deal of AI needed if trying to make sense of what the AI ‘sees’, but what if “making sense of the world” is not as important or effective as simple “collision avoidance”?

On topic, I guess whether or not an M-series Extreme would be needed is based on how they plan to implement the feature. I shouldn’t have assumed Tesla to be “state-of-the-art” in the tech.
 
Excellent point. Does make me wonder why they switched to a more power hungry chip that would actually affect range?

I currently own a ‘non-self-driving” Kia K5 that is effectively self driving for as much as I need it to be on highways. I wonder what level of hardware is currently used on those? I can understand a good deal of AI needed if trying to make sense of what the AI ‘sees’, but what if “making sense of the world” is not as important or effective as simple “collision avoidance”?

On topic, I guess whether or not an M-series Extreme would be needed is based on how they plan to implement the feature. I shouldn’t have assumed Tesla to be “state-of-the-art” in the tech.
Pretty sure Kia uses MobilEye.

Most automotive automative systems don't guarantee 100% success rate for collision avoidance.
 
Max Tech has commented that their testing shows that the GPU portion of the M1 family does not scale very well as you add cores, with the Ultra performing many tasks barely faster than the Max. They believe the issue is the size of the TLB (Translation Lookaside Buffer), though their conclusions have been criticized by some with knowledge about GPU architectures and programming.

Their latest video suggests that if the GPU of an Ultra is constrained, doubling the cores would offer little to no improvement and if Apple is asking many thousands for such an option with that constraint, nobody would bother ordering it and Apple could see PR blow-back if such a GPU falls well-behind the AMD options available on the Intel Mac Pro.
Downvoted for being objective and reporting statements not made by you lmao. @jdb8167
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.