M4 Mac Announcements

Then why not do it last year? If it’s a new die then they could presumably do it for the M4.
I dunno. It depends on when they started developing it? It could be a new low volume "Professional" chip line and therefore behind the A- and M-series chips. Intel used to do that - professional server/workstation chips based on a microarchitecture wouldn't come out until a year or more after the consumer variants were shipping (e.g. I'm pretty sure Ice Lake mobile started shipping at least a year before Ice Lake Xeon - though I think Rocket Lake Xeons were comparable in time to Rocket Lake desktop? Intel was always a confusing mess so maybe not the best example). In this way it could also have differences between it and the consumer models, like TB 5, faster cores, etc ... but the CPU/GPU cores themselves still be based on the older microarchitecture.

Keep in mind, I'm not sure I buy this explanation.
 
Last edited:
I dunno. It depends on when they started developing it? It could be a new low volume "Professional" chip line and therefore behind the A- and M-series chips. Intel used to do that - professional server/workstation chips based on a microarchitecture wouldn't come out until a year or more after the consumer variants were shipping (e.g. I'm pretty sure Ice Lake mobile started shipping at least a year before Ice Lake Xeon - though I think Rocket Lake Xeons were comparable in time to Rocket Lake desktop? Intel was always a confusing mess so maybe not the best example). In this way it could also have differences between it and the consumer models, like TB 5, faster cores, etc ... but the CPU/GPU cores themselves still be based on the older microarchitecture.

Keep in mind, I'm not sure I buy this explanation.
I’m just not sure why it would be necessary for the M3 but not the M1 or M2. New die I guess, but I just can’t believe they would allow the Max based studio to outclass the Uktra based one.
 
According to GB6, M4 Max is about 20% higher in MC than M2 Ultra. The M2 is 12P/8E, vs the M4 at 12P/4E, but the clock on M4 is also about 20% faster (base). In Metal, M4 Max is 83% of M2 Ultra, with slightly more than half the GPU core compliment.

GPU is where the heaviest work is done. The most sensible approach from a high-performance perspective (especially considering that "AI" is all the rage) would be a M4 Max Plus, which would be a GPU block extension of the regular Max, without the rest of the die on the far end. The GPU block is where the memory controllers and last-level cache are, so a Max Plus could easily take the place of an Ultra.
 
GPU is where the heaviest work is done. The most sensible approach from a high-performance perspective (especially considering that "AI" is all the rage) would be a M4 Max Plus, which would be a GPU block extension of the regular Max, without the rest of the die on the far end. The GPU block is where the memory controllers and last-level cache are, so a Max Plus could easily take the place of an Ultra.
That makes sense to me. From what I recall, your "M4 Max Plus" is essentially a Hidra, right? Then they could distinguish the MP by offering it with a dual-Hidra.
To me, the biggest naysayer would be Thunderbolt 5. There’s no way the top Mac Studio can not have TB5. And if they put TB5 on it, why not use the nearly year old M4 cores at least. M5 cores are already in production.
Yeah, I'm not sure if they can put TB5 onto the M3.

And adding to the general chorus of "it doesn't make sense" (for the Ultra Studio to be an M3): It makes no sense to have the new release of their top-end device be a generation behind every other device they've released (!)---particularly since they've made Studio Ultra customers wait nearly two years for it.
 
What we need is a chip at near reticle limit that is the CPU/NPU/etc. and another near reticle limit chip that is the GPU/etc.; with both having UltraFusion connections; this would be the Mn Extreme...?

Or maybe something based on 3nm chiplets with a 5nm controller & I/O chiplet; so Apple can really load up on the GPU cores...!

Whatever the configuration, whatever Apple decides to call it, I just hope we can see it in an all-new Mac Pro Cube... ;^p
 
T6032 and T6033 is probably the M3 Ultra. 600GB/s and 800GB/s

These could just 2x M3 Max or if its a new die something else.

This is a really good point. There were M3 Apple chip IDs that never got used, though I think it’ll be 33 and 34 right? Because there were two Maxes in the end?

So yeah we may really be looking at the release of M3 Ultras ... huh ... weird. They may not call them M3s but ... that's the family of chips they may belong to ...

 
Last edited:
The thing I’m most disappointed with, is the fact that they reintroduced the choice we used to have to make with Intel Mac Pros. You couldn’t have high end multi core performance and the best single core performance. Apple silicon allowed us to have both for two generations now we’re back to the bad old ways. A real shame and I think, a mistake.
 
$14K for a top-end M3 Ultra Studio maxed out with half a TB RAM and 16TB storage. That is not priced for volume sales. I suspect it will be a really big jump to the high-end M5 Studio/Pro.
 
1741187036548.jpeg
 
M1 Ultra took 16 months after the release of the M1. The M2 Ultra took a year. So the M3 Ultra taking 16 months suddenly doesn't seem so crazy ... it's just that the M4 came on so fast and it appearing that the regular set of M-series chips will be updated every year, makes these Pro chips look behind now - especially if the M5 (Ultra) gets released soon.


If so, which seems very plausible, that'll be really interesting ...
 
Back
Top