You're probably right, but a nice, quiet update that addresses the unimpressive GPU scaling of the Ultra and fixes the noise issues that have been reported by many users would be welcome.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
When are we expecting M2 Ultra/Max?
- Thread starter StuBeck
- Start date
- Sort by reaction score
FWIW, if Apple can produce an M3 Mac Pro with hardware RT in Spring 2023, it may be better for them to wait for that instead introducing an M2 Mac Pro in Fall 2022. And some pro apps are still in the process of switching over to AS, so Intel Mac Pro owners may not mind waiting a bit longer before moving their workflows to AS.
Apple Silicon Developer Transition Kit (2020) was introduced on June 22, 2020. That was 28 months ago. How long does it take to port from Intel to Apple chips?
When I said "in the process of switching over", I was thinking of apps for which there are native ports, but their performance and functionality hasn't yet been optimized. Each user needs to consider the apps they use, and decide whether the limitations (if any) would be a problem for them.
Sometimes the issue is that replacements for Intel libraries aren't yet available, or are available but not as performant as the Intel versions. For instance, Wolfram released a native port for Mathematica in July 2021, yet its performance for certain numerical tasks is still slower on AS than Intel, likely because the numerical libraries available to AS aren't yet as performant as the Intel Math Kernel Library.
You can also find library issues with video processing apps, like Cinema 4D's native AS port:
Last edited:
Probably more like
Mac Studio = Mn Max-D Mn Ultra (dui Max-Desktop)
Mac Pro = Mn Ultra / Mn Quad/Extreme
[ iMac Pro = Mn Max-D ]
Different chiplet factorization and the Mn Max just stays in the laptops and possibly 'skinny' iMacs.
Minimally the Ultra and the Extreme detach from the MBP laptop package ( regular "Max" ) . that laptop die is dubious as a useful building block past two. Would get better economies of scale if could spread the dies for the Mac Pro over but the Studio (and perhaps a high end iMac ).
I think Gurman has a bobby horse of trying to bring the large screen iMac back. so killing off the Studio makes hjis "new iMacs coming" theory ring more 'true' .
It can't possibly be the "Mac Pro" and the Mac Studio have high overlap with if the Mac Pro starts at $5K (or somewhat likely $6+ K ). Apple themselves stated that the Studio was a replacement for the iMac 27". if the new Mac Pro blows away both the 27" iMac and the current Mac Pro 2019 space that would be a huge 'ask' for just one product. The Studio starts at $1,999 ; that is a $3+ K gap between that and likely Mac Pro starting point. More than big enough to segment the products.
[ There are some overlap areas. if the Studio and new MP both have a Ultra then if configure up the Studio to full Ultra and 4TB drive get close to the MP starting point. But if start the Mac Pro with full Ultra and a base of 2-4TB of SSD .. can raise the MP 'floor' so not as much of an overlap. And if selling an Ultra SoC either way ... not much different than an M1 weather MBA , MBP 13" , 24" iMac situation. ]
Tossing a Mn Pro into a Mini isn't going to cover the space the Studio Ultra is in either.
If Apple dumps the Mac Studio it would more likely be to bring back the large screen IMac. Not that Apple dragged the Mac Pro back down to a $1,999-2,399 price point again.
There will be some substantive product differentiation issues if the Mac Pro is the "one slot wonder" that has been rumored at bit. If it only had some relatively crippled slot ( e.g., x4 PCI-e v4 worth of bandwidth ), then there would not be much of a gap between it and the Mac Studio when it came to I/O. But if it has two (or three) x16 PCI-e v4 full length , full height slots ( 6 pin aux power connector(s) ) , internal SATA and USB dongle connectors (like MP 2019 ) and fit the J2i drive sleds (and the similar brackets ) there would be a large 'value' gap between a "Ultra Studio" and an "Ultra Mac Pro". And they wouldn't be in the same price range. [ Likewise if have a 'rack' Mac Pro version would 'play' in vastly different space than a Studio chassis. And even bigger gap if Apple provisioned capabiltiy of putting a "compute" GPU (e.g., AMD MI210) in there. ]
Gurman's older Mac Pro rumors were about a "half sized" Mac Pro. If that is "half the volume" ( as opposed to something constrained to Apple's literal desktop footprint constraints of a 8"x8" square) then Mac Pro would be substantive bigger (and extremely likely more expensive) than the Studio.
If Apple were to 'shrink' the Mac Pro to just slightly bigger than than a Studio ( e.g. one 1/2-3/4 length slot and half height slot ) then it really would not be the "Mac Pro " pushing the Mac Studio out of the line up. It would more so be Apple 'upgrading' the Studio system with the old "Mac Pro" name ( pragmatically dropping the legacy notion of a Mac Pro altogether. MP 2013 design mantra is the new 'king' . ). Putting the "Mac Pro" name on a 'taller Mini" chassis probably isn't going to work long term with the current Mac Pro user base. The likely only way the Mac Pro 'replaces' the Studio is if Apple has designed a 'bad' Mac Pro. Perhaps, he is covertly trying to say the implementation is not really a Mac Pro.
Really would have to see a market breakdown of just how many Studio Max versus Studio 'full' Ultras Apple is selling or not. If the vast bulk are Max and binned Ultra SoCs then really not much of a 'problem'. Gurman is inventing a 'problem' that isn't there.
Last edited:
Sam, my comment about the "unimpressive GPU scaling of the Ultra" is a personal opinion. It's based on the fact that some of Apple's early maketing hype about the Ultra's GPU performance was just that, hype.
Why the M1 Ultra was doomed to disappoint | Digital Trends
Apple said the M1 Ultra is more powerful than Nvidia's RTX 3090. It was always a ridiculous comparison, and now we know that Apple's claims weren't true.
www.digitaltrends.com
And the fact that, while the Max to Ultra CPU scaling is impressive, the scaling on the GPU side leaves something to be desired (at least to me).
M1 Ultra vs M1 Max: Twice as nice, but not always twice as fast
How the M1 Ultra in the Mac Studio impacts pro users' real-world workflows.
www.macworld.com
Regarding the noise issue, it's been addressed pretty thoroughly in this thead:
there is a chance the M2 Max that will go into the MBP 14"/16" updates isn't going to go into Mac Studio or Mac Pro. Apple drops the UltraFusion connect that went completely unused in millions and millions of those systems for a die that just doesn't have it at all .
that there is a different "Max-Desktop" that is roughly a similar size but structure to scale (and disaggreate i/O) substantively differently.
So the laptops getting a speed bump "real soon now" would not reflect some new SoCs for the upper end desktop systems (inclusing the Mac Studio). The Mini could have gotten a M1 Pro so sliding in a M2 Pro should be no problem.
Apple used the M1 Max to put some volume economies of scale to make the first gen Mac Studio work. For this next gen there is a decent chance that Apple is going to need to use the now established Mac Studio volume to create a foundation to put some economies o scale to make the Mac Pro SoCs work. The higher end desktops will bootstrap their own building blocks for SoC that scale . The laptops don't need any "multiple dies in a Package" scaling at all. So stuffing that overhead in them is actually a waste of wafer space (and money).
the Ultra SoC isn't 'cheap' , but appears to be selling in some very decent numbers. If combine Mac Studio unit numbers with some decent, but much smaller Mac Pro unit numbers (at even higher price points), there is decent chance there is enough profit margin lift to offset the hit of jumping off the "Max" MBP 14"/16" bandwagon.
Apple tossing the M2 Pro into a Mini and perhaps a M2 Max into 'skinny' large screen iMac chassis they could sell more 'plain "M2 Max" would 'lose if Studio dropped them on next iteration in Summer/Fall 2023. Similar Apple could boost the desktop "Max"/"Ulra" SoC if tossed it into iMac pro chassis (at a decent price to generate volume).
Apple is spec bumping the MBP 14"/16" in 12 months, but have left the Mini drfting on the M1 for almost two years. The M1 24" iMac is rumored to be drifting well into 2023. Apple's actions so far is that the desktop Macs are on 'slow motion' update cycles. Probably not going to be on the same update cadence as the laptops (that Apple probably cares more about because they generation most of the revenue and profits. ). Likely, we will not going to see "Ultra" and "Extreme" SoCs come out at the same pace as the plain , mono-die packages.
IMHO a M2 "Max sized " building block made on TSMC N3 would make more sense as the desktop SoC building block than whatever they are going to do with the latpop version (TSMC N4 ? Or worse N5P ? ) . If so those SoC are sliding into 2023.
Sam, my comment about the "unimpressive GPU scaling of the Ultra" is a personal opinion. It's based on the fact that some of Apple's early maketing hype about the Ultra's GPU performance was just that, hype.
Why the M1 Ultra was doomed to disappoint | Digital Trends
Apple said the M1 Ultra is more powerful than Nvidia's RTX 3090. It was always a ridiculous comparison, and now we know that Apple's claims weren't true.
And the fact that, while the Max to Ultra CPU scaling is impressive, the scaling on the GPU side leaves something to be desired (at least to me).
Apple spent 2-3 sessions at WWDC 2022 on how to do better optimized scaled GPU code with new much improved profiling/debugging tools . For better or worse that was a couple of months after the Studio shipped. The 3090 comparison was an obvious cherry picked. ( not like Intel, AMD , and Nvidia aren't doing the very similar things on stage. ). but there are substantive software gaps during the first 3-4 months of the M1 Mac Studio's service lifecycle. Pretty good chance it won't robustly shine well until a couple of months after macOS 13 (Ventura ) rolls out.
The other "perfect scaling" issue hype train was driven at least as much ( if not more) but folks chattering in forums like Mac Rumors. I'm not sure why folks deeply bought into that. UltraFusion makes the 'speed bump' relatively very small going from die to die ... but still going die to die. any portion that isn't "extremely embarrasingly parallel" and likely to see hiccups. Plus Apple is going to be running huertics trying to keep work grouped to save power , but those heuristics largely only operating in "secret lab" before launch. [ Somebody at Apple thought the power settings on the MBP 16" 2019 were set right when the shipped it out the door too after "failure to use imagination": testing. ]
The experiments where folks jam microphones into the back of Mini/Studio/Mac Pro where no user sits are just bad experimental design. Almost nobody normal sits with the back of those systems facing the user. If put the microphones where a normal ears would actually be , then not going to get the same results.
as for the power supply 'coil whine' ... it took Apple how many years to fix the butterfly keyboard? Killing the Studio "gen 1" much shorter than its planned lifecycle probably isn't going to happen. Also probably is not a relatively short planned lifecycle either. If it is a quality control in the supply chain maybe that gets fixed before a model upgrade but it isn't like Apple has tons of alternative suppliers over the short term for highly non commodity parts.
I bought a MBP14 M1 Max this week, I had been told by numerous apple store staff the M2 update would be marginal, similar to the iphone 13 to 14 revision.
At a certain point, I guess you have to just bite the bullet because there will always be new tech being announced/marketed and that feeling of "if I just wait another few months to get the latest" becomes self-defeating. I've been buying computers since the Ti99 4/A since about 1978 and learned that there will always be something new coming around the corner...
At first I was thinking that cancelling the MS makes no sense; they just filled a middle market need with a desktop unit that users including me have been clamoring for like 7 years for. Basically, there are 3 market segments:
1-sub -$1,000 budget buyers apple wants to pull away from the Windows universe
2-mid-market/prosumer buyers with budgets up to about $2,500 who might play games, use photo/video editing software but not professionally, and want more power and speed than the mac mini can offer
3-professionals right up to Hollywood VFX houses who need massive computing power and can afford $25K on a machine and will buy ten of them to daisy chain together to render the next Avengers spec EFX
The MS nails the 2nd category, but then I was thinking about the laptop market which apple has bisected into just two categories, so maybe that is where they are moving towards; a unit for the bottom that can be specced up to a pretty powerful machine with upgrades, and a pro machine that starts with a $5-$6K price tag but can upgraded to $40/$50K with massive firepower.
I think we might be looking at this backwards; we are used to looking at the machines themselves and how they are marketed rather than having apple tell the consumer: "what kind of work are you going to use the machine for, and then select a package", without regards to product name/identity - just whether it is placed in a portable shell like a laptop or stationary box like a MS.
I see where you are coming from and I fit right in that Group 2.
Issue is that the entry point machines like the Mini and MBA are lacking capabilities that even non-pro folks want. Example - dual external display support. It's patently ridiculous that the M2 MBA can't handle two 5K displays. And while the M1 Mini can technically support two displays one has to be via HDMI and that sucks.
I guess if they end of life the Studio they will provide some sort of prosumer version of the Mini which could be fine. But it will end up being close to $2K so what's the point.
The Mac Pro design is so iconic, I hope they don't change it too much. I know the design cues of that case are more geared towards the issues with Intel processors, but still, I think the case continues to age well, unlike many other PC cases.
I personally think the new Mac Pro will be the first generation of desktop class processors based off N3. Perhaps its not called the M3 Max whatever, but X1 or X3 or something to denote something much larger.
Yeah, I'd be fine with something less than perfect GPU scaling from the Max to Ultra, and it's actually not a huge issue for me personally. It just seems that there is something "off" with the current Ultra in this regard. Maybe it'll work out over time with better software optimization, but given the 2x jump in price from Max to Ultra, it would be good to see Apple address this in an update.
I might split the 3rd group in two. In the lower tier you have a lot of photographers who are doing heavy stuff by photographic standards (like very large panos and big focus stacks where you're playing with hundreds of 250MB, or bigger, tif files). We've been working on Mac Pro's less for GPU grunt and more for the ability to add RAM.
Another consideration is that, as nice as the Apple screens are, they are glossy and don't have built-in calibration monitors. Compared to an Eizo, they really don't work for us. If Gurman is trying to nudge Apple away from headless desktops to iMac's I hope he isn't successful.
The Studio looks interesting, but when you factor in a configuration that gets you to 128GB RAM and maximum on-board storage (8TB) you have a $7,000 machine and $8,000 with the GPU bump to 64-core.
I'm certainly not in the higher end of group 3 (Hollywood) and those folks probably deserve their own category because that's where the really heavy work is going on. But, like them, I'm holding out to see what the 8,1 Mac Pro brings. It's bound to be expensive, but if it offers a compelling blend of performance and expandability (I know, that's a big "if") it might make sense to put the $7,000 to $8,000 I'd spend on a Studio toward another Pro.
We'll see.
That's true!
When it comes to needing a matte display, do the nano-texture screens not work for you? The last 27" Intel iMac had that as an option, so I assume a large AS iMac would as well. [Yeah, even if it did you'd still have the calibration issue; but I was specifically wondering whether Apple's nano-texture didn't meet the matte-screen needs of some video pros, or if you were simply unaware that it was offered with the large iMac.]