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.
Mac Pro and Pro Display XDR Launched Three Years Ago Today
- Thread starter MacRumors
- Start date
- Sort by reaction score
I take landfill to mean “not recycled”. And, even better, not returned to Apple for Apple to recycle. Anyone who’s going to recycle is going to recycle. Anyone who isn’t… isn’t. For person that isn’t, all they’re throwing out with today’s Macs is the Mac they bought, not the Mac plus the several swapped out RAM upgrades, storage upgrades, GPU upgrades, etc.
IF the GPU and RAM is being responsibly recycled, not much. But, that’s not the folks of concern. The folks of concern are those that, when they swap out the GPU, that old GPU is now in the landfill STARTING that person’s impact on the environment while they’re still using the computer! Then their next GPU, their next sticks of RAM, their next storage, all going into some landfill while the person’s using the computer. For Apple’s current systems, if that person is a dumper, their system doesn’t start impacting the environment until they actually dump it.
Yes, proprietary GPU’s are good at proprietary GPU workloads, that’s true for Nvidia and Apple both. And, it also doesn’t change the fact that a lot of people will continue to have difficulty grasping how different Apple’s solution is and how it’s not suited for eGPU’s. It’s been true since June 22, 2020, before the first M1 system even shipped, but there’s a decent number of folks that still don’t get it.
Apple’s just focusing on the Pro’s that want what they’re selling. It’s not too different from the way Apple treated the entrenched FCP7 Pro’s that wanted more of the same while Apple wanted to go in a different direction. Yes, a lot of folks dropped FCP and never went back, but as of some number of years ago, there are more active users of FCPX than there ever was of FCP7.
I’d think that most of the “serious pro business” folks that aren’t that fond of macOS and not tied to Apple only solutions like FCP moved on years ago. I actually helped migrate a few… I think it was a few years before the current Mac Pro came out? Most of them stopped paying attention to Apple awhile back, but some saw the new Mac Pro and it’s price and contacted me to let me know that they didn’t have to care about that anymore
(I was actually a little glad, too, because I had told them that Apple was done making the kind of systems they care about and was relieved when Apple showed that to be true! I knew there was a small risk I might be migrating some folks BACK, but that never came to pass).
How I feel about it all...and ill be honest, I think I'm right
https://forums.macrumors.com/threads/mac-pro-8-1-what-if.2368905/
What the 8.1 will REALLY be
https://forums.macrumors.com/threads/mac-pro-8-1-what-if.2368905/
What the 8.1 will REALLY be
I had first Mac Pro then upgraded till trashcan, which I always thought was pretty cool. Now I have a Mac Studio Ultra which is overkill but the mini is underkill and I do desktops with large screens, although I do have the latest 16" MacBook pro I rarely use. I might get the next high end display, the studio display is too small for scores. For now I use a 4k thunderbolt 3 32" display. I can't imagine the new Mac Pro costing less than 6k. It would be super overkill for what I do nowadays... but maybe I'll go get a phD or something.
Apple pragmatically really hasn't flip flopped that much.
MP 2013 --> iMac Pro (2017) --> Mac Studio (2022 ) [ not 4 years for last, but also not global pandemic during the first two. The M1 Max used in a Studio was in 2021 ... 4 years. ]
[ Mini 2014 --> mini 2018 . It isn't just the Mac Pro that has slow motion iterations. Mini 2020 -> 2023 ? ]
They never completely abandoned that ~400W , literal desktop with Mac Mini sized desktop footprint constraint.
There was a long gap between 2012 and 2019 but there is a sizable divergence there also. The entry level price increased 100% . Hard to say that is the 'same' target market for the product.
There is little indication that large bulk of the high end buyers move at anything like a "every 12 month" cadence like many of the iOS/iPad iDevices do. Or 18 month cadence.
There is lots of commentary in Mac Pro forum about some Power G4 that is still doing work in a editing bay . Or how another round of component stuffing into a Mac Pro 2010-2012 was going to work out OK . Or buying a very high priced Mac Pro to write down on taxes (which for USA is a more than two year schedule ). Or how Apple "has to" offer support for the Mac Pro for the next ten years ( which Apple doesn't outline at all in their policies. )
Pretty good chance that Apple presumes that most folks who bought a $8-20K Mac Pro in 2019 are likely to sit and squat that machine for 5-6 years. The even higher entry price means there is an even high "sunk cost" preception attached to it. That typically invokes a desire to keep the system longer. Sell a substantive number of those folks some > $5K W6xxx upgrades from Vega Pro II and an even bigger bunch on the 5-6 year cycle time.
In short, pretty likley the average Mac Pro buy average lifecycle utilization window is long than that of the average MacBook Air or Mini buyers. So why should cycle through refreshes at the same rate if the uptick isn't going to be there. [ Yes, some buyers will be coming out of their 4-6 cycle at the 'wrong time' given in 2016,2017,2018, 2019 had folks going in at yearly pace and long gap after 2012. ]
It isn't Apple as much as the customers who clamour for the every 12 month Mac Pro. Every Jan-April there is a huge uptick of "Apple's going to present a new Mac Pro at WWDC <that year> ). It is appears like a broken analog clock every year. Pretty good chance that Apple presumes that is going to happen (as much as paying attention to the chatter about the Mac Pro out in the forums ).
2019 + 4 is 2023. Which looks like they are going to hit. Pretty good chance they were aiming for very end of 2022 which would have been three instead of four. Something closer to 2-3 years is probably were they are drifting toward once finish with transition work. The big packages SoC are not likely to iterate very fast.
In directly, they have essentially said so. The explicitly claimed that the Mac Studio 'finished' the iMac 27" transition. That basically implicates that it is not the Mac Pro transition product. And they initially claimed to transition all of the Macs in "about two years". So the Mac Pro is "undone". Sure they have tagged the Mini as 'done' and are still selling the 2018 Intel model going into 2023. ( 5 years ! )
It would be specularly lame to say "oops , while the Studio wasn't a replacement last year ... we added a few more GPU cores and now it is" .... there isn't even a Steve Jobs Reality Distortion Field large enough to wrap about that junk.
Apple is probably going to release something that has some PCI-e slots in it. May not have DIMMs or 3rd party GPUs but there is no rational reason not to do a system that can take at least some of the 50 PCI-e cards that ALREADY WORK with macOS on M-series and deliver a system so can have those inside. They have already laid the software ground work to enable it. Then walk away? Really?
If they just refactor the chiplet disaggration so they can put on one or two x16 PCI-e v4 lanes complex(es) they'd have enough to provision a couple of slots. Instead of having way to many TB controllers they can't really effective use (if scale past two , overly fat , laptop SoC used as 'chiplets'. The M1 Max is not good chiplet design. It is cost effective if not going to ship a Mac Pro until as late as possible, but do a Studio in 2021. But for M2 or M3 and later they need something way, way ,way better designed to be a chiplet. )
There have already been a sigting of a one-slot-wonder and six slot Mac Pro. The likelihood that it has zero slots is pretty low. It isn't going to take a huge amount of re-design work to swap out 4-8 TB controllers for a 1-2 PCI-e compleses. Even if they didn't and had way too many TB complexes, it is easy (and relatively cheap) hackery just to slap some discrete peripheral TB controllers inside the box and provision the PCI-e slots out of those extra x4 PCI-e v3 allocations.
Apple doesn't need another Studio named "Mac Pro" to compete against the 'regular' Mac Studio. They need something different in the line up. Even if both share soldered LPDDR5 RAM and SoC packages.
I suspect Apple really didn't know when. They had the outlines of the Mac Pro they wanted to do , but didn't know when could make the SoC packages to go inside the box. So just 'later" ( as in longer than 6 months from now ).
At this point though, they should know what a plausible timeline is. Drifting into 2023 after basically implying that they'd finish before 2023 deserves some "dog ate my homework" explicit excuse. Even if it is a "new Mac Pro not coming in 2022 " ( which is pretty much obvious at this point. The number of these kinds of tech stories os 'Apple flaked on 'about two years' is only going to increasing after the New Year starts. )
if they are going to kill off 3rd Party display GPUs then just 'rip the band-aid' on that too now. They have done nothing for over 2.5 years at this point ( in terms of driver development) . So it really isn't "new news" , just whether it is the normal status quo going forward or not. Putting people looking for a newer container to throw their 3rd Party GPU into a holding pattern circling the "Apple Airport" when you are never going to let them land ... what is the point of that. Send them to another airport.
If it is just grossly delayed down the 'to do' list and a whole category they'll eventually get to ... just say that. They don't have to name specific GPUs package(s). Just that the general class will get addressed. When Apple announced that DriverKit was replacing IOKit (and kernel extensions (kext) were deprecated and going away completely eventually) they said they'd be flushing out the DriverKit classes/categories to replace the IOKIt classes. Only they never did address the GPU class of IOKit. Deliberate or just lots harder than they thought?
What? I’m not the biggest Nvidia fan myself but facts are facts and for professional workflows an Nvidia GPU would have been ideal because of CUDA and Nvidia have the best-performing graphics cards in professional workflows and in gaming. Also didn't AMD have better-performing CPU options than Intel at that time?3 years old and still the most powerful 3D and other GPU hungry machine has produced to date. And it's not even close. A fantastic machine for anything GPU intensive and you cannot get better from Apple in that regard.
Let's see what they do with their own AS in the next Pro
Look at the problems Apple has had fulfilling orders for build-to-order machines. Personally, I am ok with soldered RAM. I think the storage should be upgradeable and that would at least make the manufacturing less complicated and allow them flexibility by just being able to pop in 1-8TB (or more in the future) into the machine. Would be better for customers too.
True, if you take that as the real "Mac Pro" lineage. For people who needed internal PCIe and storage it's still 2013 (March 2013 in Europe) to late 2019/early 2020 with no shipping product. And even headless desktop (trashcan) to all-in-one (iMac Pro) (by which stage the Trashcan was badly outdated) and then back (Studio) is something of a flip-flop.
Nothing between 2012 and 2019 had expandable internal storage or PCIe GPUs cards without using an external Thunderbolt enclosure & the subsequent limitations in bandwidth (20 Gbps until the iMac Pro added TB3). Anybody who needed that was the target market for the 2019 MP. I believe quite a lot of people had been keeping their 2010-12 model Mac Pro towers alive until then.
I don't think any "serious callers" expect or would even like an all-new Mac Pro every 12 months, but 4-6 years - by which time even an upgradeable machine will be falling behind on PCIe/DDR/USB/Thunderbolt versions - is a bit much. There's also the distinction between incremental, backwards-compatible upgrades and "exciting" new form factors which affect the workflow. We're not just talking about individuals upgrading their personal machines (with a keen eye on rumours and product cycles) For a larger business you'll have to kit out new employees, replace failed equipment etc. as and when the need arises - and even your main equipment-buying schedule might be dictated by accounting and leasing issues. Neither stuck with buying end-of-lifecycle kit or being forced into workflow-altering format changes (Tower to Trashcan to iMac to Tower) is a great choice.
If you rule out GPUs then most of the remaining cards with M1 support are probably mostly x4 cards that work happily in a TB3/4 enclosure, and Mac Studios (particularly ultra) have significantly better TB bandwidth than the old Trashcan. Plus, with ASi MacBook Pros now much closer to desktop Macs in performance, TB lets you easily swap your specialist PCIe cards with individual's laptops.
...but that brings us back to the elephant in the room: using external GPUs nixes many of Apple Silicon's performance benefits of unified RAM and on-chip GPUs/Codecs etc. and - for applications where power consumption isn't paramount - you're back to a PC tower that can't run x86 code.
If you don't want external GPUs then yes it probably isn't rocket science to turn some of the spare TB4 ports on an Ultra or Extreme chip into PCIe. If all else fails you could use an existing chip with TB peripheral controllers to get the PCIe back, and that's assuming the Apple Silicon on-chip TB ports can't already be configured as PCIex4 (at the end of the day, physically, it's just 4 pairs of fast serial lines). I'm not saying that an ASi with PCIe is impossible - just that, without GPU support (which nixes Unified RAM) it may be time to revisit the idea of external PCIe expansion.
Sure they could - they just wouldn't spin it like that, any more than "Heck, we can't afford to give away 5k screens with iMacs any more and still afford designer door knobs for our HQ, so here's a fat Mac Mini and an over-engineered 5k screen with a the mark-up we deserve".
Something like "The Mac Studio has been a magical success and people love the neat, courageous, compact design, so today we bring you the Mac [Studio] Pro Ludicrous Extreme with even more power... why would anybody want a tower PC? Can't innovate any more my ***!" - and the crowd goes wild.
Personally, I think a 1U rackmount Mx Ultra machine (with matching PCIe and external storage units available) would be a more realistic application of the Apple Silicon tech we've seen so far & if I wanted a big box'o'slots with full-size GPUs I'd go with AMD x86...
So how does that translate to market share? The last 10 years must have seen a huge surge in semi-pro video creation for YouTube etc. (and even humorous cat videos now need to be in 4k HDR
) as well as a boom in TV production driven by streaming services, so I'd expect the whole video editing market to have boomed. I'd have thought that the Mac Studio concept was spot on for serious YouTubers, and targeting that rather than more traditional studios would be a good business plan.
Certainly not publicly - that would damage Apple's business interests. Apple does have large business/commercial customers where some aspects of product plans would be released under NDA.
I love my Studio Display XDR, purchased mine used on eBay and saved about $1200. It has worked flawlessly for me, and I cannot see ever going down in resolution from 6k when talking about the 32" screen size.
I really don't see a future for Mac Pro like machines any more.
The pool of users who need uber high-end computers and who still use OSX has to be so small as to make the investment not worth it.
The Mac Studio doesn't have any problems for rendering 4K and 8K video or extensive audio production, so what's left?
The pool of users who need uber high-end computers and who still use OSX has to be so small as to make the investment not worth it.
The Mac Studio doesn't have any problems for rendering 4K and 8K video or extensive audio production, so what's left?
Most of us don't need Mac Pro level power - but there are applications out there in 3D and scientific computing that really do need insane amounts of RAM and PCIe bandwidth. The 2019 Mac Pro feels like it was designed and priced for the likes of Pixar, and until the Studio arrived that left a bit of a gulf where the old $2000-$4000 cheesegraters and trashcans used to fit.
I found the videos from Neil Parfitt gave a good "reality check" on why some people need something like the Mac Pro even for things like music composition (let alone 8K video or 3D). TLDNR is that his workflow needs a lot of instrument samples loaded into RAM and needs PCIe for multiple specialist audio I/O cards. This is a link to the end of one video where he shows the Trashcan-based system that his 2019 Mac Pro was replacing:
OTOH, I suspect people like him would be quite happy for another 5 years of support for the x86 based Mac Pro before embarking on their Apple Silicon adventure. There's a newer video explaining how he tried a Mac Studio as a "mobile" system and it Just Didn't Work for him.
Plus, the guy has a flux capacitor module in his insane Eurorack synth setiup
!I had a pretty different experience at the studios I worked at. I worked in far fewer than 50, and perhaps we are in a slightly different industry (I work in animation). But the Mac Pro 2008-2012's were ubiquituous in all the studios I worked at. Upon Apple upgrading to the 2013, studios beta tested it but its lack of configurability prevented it from being rolled out widespread to the entire studio. Specific things like large amounts of storage, ram, interface cards (such as fiber networking) or better video cards were commonly rolled out in differing configurations to different users based on need. The 2013 Mac Pro could not service most of these needs.
At this time the IT depts also had to compare price/performance against the competition, such as Linux/Windows based HP Z workstations. The multiple GPU upgrade setup that was not sufficiently upgraded over 6 years effectively killed its adoption in studios I worked at. While some in this thread pointed to thunderbolt e-GPU solutions, those made the setup even more expensive, lacked the full throughput of an actual PCI card, and as far as I'm aware Apple never supported using Nvidia cards on Mac OS with driver support, which had many features that were desirable for 3D workstation applications - neither the Quadro or Geforce line was ever fully supported unless you went over to the Boot Camp side of things.
By the time the 2019 Mac Pro rolled around the scene with its fancy wheels, 6 years after the 2013 Mac Pro, the transition to PC workstations at all the studios I worked at that previously had Mac Pros was complete. IT depts had moved entire companies away from Macs at that point and it was too little, too late for the new Mac Pro configuration to keep these studios on their platform.
The way I see it, its always been an aspirational product line - the top end for the 'Pro's' - it defines what the platform is capable of. If there isn't a serious machine to run programs natively such Nuke/Katana, or any of the Autodesk suite (which still only run under Rosetta), then those high-end creative apps just won't be on Apple Silicon Mac OS. And fewer high-end software packages for the platform leads to creative brain drain to other platforms that do support it.
I think reliability is a key aspect of things that a lot of people don't really focus on. If you're paying this level of money, its because its mission critical and is making you money. If one of the main pieces of your hardware workflow is unreliable its a total non-starter. Why you would ever want to pay more for an external solution that can be accidentally unplugged, has lower bandwidth, and more constrained cooling is beyond me.
This really nails it as well. A huge factor in the Mac Pro is reliability. I just need all the components to work and not have any disruption in work. Just turn on the computer and it powers/activates all the cards/drives/GPUs/etc.
Reliability is more important than ultimate speed and power for me. Fortunately my Mac Pro has paid itself many times over already through produced work with it, and a large part of that is its consistent reliability.
I LOVE AMD and quite frankly, I have a $12k puget system sitting in a cart with 2 RTX 4090's in it that I've been debating all week if I want it sitting next to my $50k Mac Pro. Look at my signature to see it's internals...I am telling you first hand, my machine is approximately equivalent to 3 RTX 3090's when it comes to realtime workflow in OctaneX in Cinema 4D. I been using this workflow for my clients for a long time and both realtime work and rendering and on export, it simply flies.
Same cannot be said for Adobe products since Adobe refuses to rewrite their software to take advantage of todays' systems and to not crash every 5 seconds...but I digress. I use this machine strictly for 3D animation and VFX for film and television. And I use it because my entire ecosystem "both at home and at the production company I own", are 100% Apple based.
But I think you didn't read what I wrote properly...I said this is the most powerful and GPU hungry machine APPLE has ever made...this has nothing to do with Nvidia. That wasn't my comment at all lol.
Sadly I agree, but it doesn’t change my statement. Apple made an apology about the trashcan, then touted “modular design” and “upgradability”. Then, right at the time where the first movers might start to think about actually upgrading, you shift to a different platform… Apple (obviously) knew when they launched Mac Pro that Apple Silicon was coming. Touting upgradability in a platform they knew they were going to trash five years later is Samsung-level shithousery.