Apple doesn't have $3T. Apple stockholders have $3T. Apple not equal Apple stockholders. And extremely likely not going to find most Apple stockholders excited about some moon rocket boondoggle that just means their money down the drain.
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.
Apple Executive Discusses New Mac Pro's Lack of Graphics Card Support
- Thread starter MacRumors
- Start date
- Sort by reaction score
Apple doesn't have $3T. Apple stockholders have $3T. Apple not equal Apple stockholders. And extremely likely not going to find most Apple stockholders excited about some moon rocket boondoggle that just means their money down the drain.
I waited for new Mac Pro, in case it would be significantly faster than Studio. But it is not.
All I ever wanted for my project studio was a Mac with AS, with a CPU comparable with i7/i9 and at least 128GB Ram. In a small enclosure, quiet under load (not only at idle).
M2 Ultra Mac Studio is all that for me and (I dare to guess), for 90% of pro, project and home music studios.
You dont reach a market cap of $3 trillion by pouring R&D money into a product line that would likely sell about 75,000 units annually.
You do that by burning R&D resources into a product line that ships nearly a quarter billion units annually.
I was looking forward M2 Extreme for purely nerding out satisfaction.
But odds are there were fab limitations that could not be circumvented before June.
So we'll have to wait for 2025 for it to show up.
In 3 months time I look forward to an iPhone chip with Ray Tracing cores as it would mean the M3 will get em by 2024.
Yeah. I keep my studio computers for 3-5 years usually. So by then, it will be M4 or M5 with 512-768GB of Ram and hardware RT 😏
I’m presently rocking 2018 Intel Mac Mini (bought in 2020 to ride the wave of AS transition). It will be quite an upgrade for me.
Weird that I keep seeing videos of real time effects at full resolution (no downscaling) by 3D designers repeatedly mumbling wow, I’ve never been able to do that. That was on the M1 Studio.
It’s almost as if Apples tile based GPU strategy works remarkably well when the software is actually designed to use it…
This is one of the reviews I was talking about: Zbrush
At 13 minutes in he tries to break it by upping it to 300+ million polygons (“I can’t imagine why I’d ever have to do this”) and then he’s genuinely surprised he can still sculpt it.
This thing is powerful.
It’s almost as if Apples tile based GPU strategy works remarkably well when the software is actually designed to use it…
This is one of the reviews I was talking about: Zbrush
At 13 minutes in he tries to break it by upping it to 300+ million polygons (“I can’t imagine why I’d ever have to do this”) and then he’s genuinely surprised he can still sculpt it.
This thing is powerful.
Last edited:
If you have watched this video below of a film/music/tv professional is your workflow/use case similar/same?
What is your point of view on it?
I am just running a project studio/small indie label, so my point of view will be slightly different than his.
I’ve had Performas, G3 Tower and G4 MDD Tower. Then I’ve had Mac minis, after Apple switched to Intel. Never had the need for PCI cards myself. I’ve always had either FireWire, USB or TB audio interfaces.
As for RAM, I think more than 128-192GB is only required for large orchestral templates.
So for majority of professional musicians, new Mac Pro is fine if they need to put Pro Tools cards into it or UAD cards. Otherwise, everyone will go with Mac Studio.
M2 Ultra is close enough to i9 in performance, IMHO.
As for 4090, who needs THAT in music studio? To heat it during winter, perhaps 🤣
The level of noise from those 3 fans alone…
If a M2 Extreme were available with 2x of everything would you have bought that?
It’s not laziness. It’s intelligent product management: Improving the experience for 80% of users by sacrificing the 20%. The integrated chip solution which disables GPU support is what enables significantly better performance for users that don’t need it.
For me, this is a business expense so I do have a finite budget 😆
I have allocated £8000 for my next studio computer update. So probably not, looking at how they’ve priced Mac Pro.
At the moment, I’m just below £7000 with planned M2 Ultra, 192GB Ram, 4TB SSD and AppleCare+.
For 75,000 units shipped annually worldwide... that's just plain smarts.
Law of diminishing marginal returns is at play here.
Have you tried running the same task on an Apple Silicon mac with less ram? What happens if you do? Honest question, I don’t use those workloads, I am a light user. I’m just curious to know if you will be unable to perform the task, or if the task just becomes slower?
I’m asking because I use an 8 gig M1, which is much faster than my 16 gig Lenovo for anything, and I have never seen a popup or felt a limitation. Unlike my Windows machine - I have files that I can’t open on it due to lack of memory (according to the error message), that I then bring to the Mac which opens them without any noticeable lag.
I don’t know how this experience scales to heavy use, but I’m curious 🙂
I don’t think sales volume matters at all. I think it is a loss leader. But if 80% of those 75.000 has a unique experience and 20% are lost, that is a much better loss leader than if 100% has a bland experience.
I think the $1k MSRP bump covers that loss leader.
May be better termed as a "lean leader" in terms of margins?
Margins are not as good as say any iPhone or M1/M2 Mac?
It amazes me that other think a Mac Pro is a gaming PC without RGB lighting or glass side panels.
Not sure which is the correct term, I’m quite sure they have their direct per-unit manufacturing costs covered. I’m questioning whether they will have their development costs covered. In my company that is a loss leader, even though you make a direct profit per unit sold, because overall the product is not profitable.
Edit: By the way, I would expect the margin to be higher than on an iPhone.
"Lean leader" to me would be it covers BoM, logistics and R&D cost but the net isn't that good to be that eager to service it.
Apple moving from Xeon to Ultra chips would likely allow them to refresh it every 15 months with the Mac Studio.
So no "high & dry" for 3.5-7 years without a refreshed Mac with PCIe slots.
Apple doesn’t do loss leaders (which is defined as “a product sold at a loss to attract customers.”). If Apple has something for sale, it’s being sold for a profit, and likely 20% or more.
My guess is that the price set for the Mac Pro includes R&D, BoM, plus how many they expect to sell over it’s useful lifetime, plus servicing after the sale. We’ve already seen that they’re getting more use out of the frame they designed for the last generation, so they’re still getting back dividends from that R&D. There’s no need to cut the price on it because anyone that needs it is going to buy it, there’s no other macOS system supplier available to go to.
Many have said Apple not offering support for nVidia 4000 series PCIe video cards on the 2023 Mac Pro is a mistake because "the world" is developing AI on nVidia 4000 series video cards. Do those development applications even have a macOS-native version? Or are they presuming that AI developers would program via command line Unix using the Terminal app?
Same with those who want an nVidia 4000 series PCIe video card to play games. I would presume all Apple Silicon-native games would be designed to work within the confines of the GPUs on the M series of SoCs so would an nVidia card really make a huge difference for them?
Same with those who want an nVidia 4000 series PCIe video card to play games. I would presume all Apple Silicon-native games would be designed to work within the confines of the GPUs on the M series of SoCs so would an nVidia card really make a huge difference for them?
Last edited:
Sorry, but that doesn’t sound like an educated guess to me, since 20% is way too low. And like I stated above, I’m 100% sure Apple makes a per-unit profit. I am questioning whether per-unit profit (which is margin, not “profit”) times number of units sold covers the overall project expense. Which includes a lot more than R&D. But, I may be too pessimistic on numbers sold.
I keeping wondering why it took such a long time to release what is essentially a Mac Studio inside a box with PCIe expansion slots that can't even support graphics cards.
I'm very disappointed with the Mac Pro.
I'm very disappointed with the Mac Pro.
The delay is possibly, if not probably, due to Apple working on making an "extreme" class SoC and offering large-capacity off-package external memory access, neither of which they could make work to their satisfaction.
So they fell back to just putting in an Apple Silicon SoC and motherboard in the existing case to meet the needs of the small portion of their user base who run Apple Silicon-optimized software that connects to non-GPU hardware that is on cards.
Unified memory access is preferable, though it can be done with external GPUs as AMD did with EPYC Trento and MI250X for supercomputers; AMD simply modified the IO die to include dGPUs in the unified memory space via Infinity Fabric. Apple simply isn’t interested. AMD’s next MI300 is more akin to a very large GPU with a CPU on-package with a ton of HBM3. Should also be noted that Apple still refuses to do business with Nvidia.
Generally, Apple designs their silicon around optimal perf/W of the silicon itself (obviously, right?). Their parts rarely exceed 3.6-3.8GHz as the V/F curve moves outside of the optimal range. I’m sure Apple could boost their big-cores to something like 4.8GHz, but due to design constraints, it’ll end up consuming much more power for a meager 3-5% gain. Why? AMD and Intel purposely dedicate transistors for higher clocks (millions of them in order to maintain optimal clock cycles even at higher speeds), while Apple does not. So, pushing beyond intended design doesn’t bring magical benefits for Apple. It’s a steep regression in perf/W instead, as critical parts of the architecture will lengthen clock cycles per instruction or operation to prevent errors; if there are multiple clock domains, parts that can’t run at 4.8GHz will stay at lower clocks, causing an imbalance in the architecture. You simply can’t force silicon designed for efficiency to perform well outside of its design parameters.
This applies to M2’s iGPU as well. It operates at 1200MHz where AMD and Nvidia are pushing over 2500MHz because extra speed aids the entire architecture, especially fixed function parts that don’t scale as well as actual compute logic; fixed function is exactly what it sounds like, doing one thing exceedingly well in hardware (something Apple uses in the form of various in-die accelerators). Things like rasterizers, ROPs, and yes, ray tracing units. To improve fixed function units, you either need more units (extra die space) or higher clocks or you can simplify the operations FF units do and offload some computation to scalable compute units at a cost of power consumption. Media codecs are also fixed function, but reside in a different area of the GPU and have their own clock domains. The same rule applies though.
Anyway, tl:dr - Dedicated external GPUs are possible within unified memory space, but Apple wants you to use its silicon instead of wasting all of the die space they dedicated for M2’s iGPU.
Generally, Apple designs their silicon around optimal perf/W of the silicon itself (obviously, right?). Their parts rarely exceed 3.6-3.8GHz as the V/F curve moves outside of the optimal range. I’m sure Apple could boost their big-cores to something like 4.8GHz, but due to design constraints, it’ll end up consuming much more power for a meager 3-5% gain. Why? AMD and Intel purposely dedicate transistors for higher clocks (millions of them in order to maintain optimal clock cycles even at higher speeds), while Apple does not. So, pushing beyond intended design doesn’t bring magical benefits for Apple. It’s a steep regression in perf/W instead, as critical parts of the architecture will lengthen clock cycles per instruction or operation to prevent errors; if there are multiple clock domains, parts that can’t run at 4.8GHz will stay at lower clocks, causing an imbalance in the architecture. You simply can’t force silicon designed for efficiency to perform well outside of its design parameters.
This applies to M2’s iGPU as well. It operates at 1200MHz where AMD and Nvidia are pushing over 2500MHz because extra speed aids the entire architecture, especially fixed function parts that don’t scale as well as actual compute logic; fixed function is exactly what it sounds like, doing one thing exceedingly well in hardware (something Apple uses in the form of various in-die accelerators). Things like rasterizers, ROPs, and yes, ray tracing units. To improve fixed function units, you either need more units (extra die space) or higher clocks or you can simplify the operations FF units do and offload some computation to scalable compute units at a cost of power consumption. Media codecs are also fixed function, but reside in a different area of the GPU and have their own clock domains. The same rule applies though.
Anyway, tl:dr - Dedicated external GPUs are possible within unified memory space, but Apple wants you to use its silicon instead of wasting all of the die space they dedicated for M2’s iGPU.