All companies, no matter how generally successful, have missteps.
Except Apple has had missteps with their last 3 "pro desktops" in that they've gone for years without updates and/or been discontinued long before the replacement is announced and/or replaced by radically different concepts that need a workflow re-think... and although #4, the 2019 Mac Pro still "is what it is" in the short term it's looking increasingly like another one-off.
It’s funny all these people saying that it’s a piece of crap.
No, it wasn't a piece of crap: it was clearly a great machine
if it met your needs and budget. The problem was, a $5000 all-in-one Xeon workstation with no choice of screen and limited internal expandability/configurability was a bit of a niche product at a price range where users are starting to have highly specialised requirements. The underlying problem was not the iMac Pro itself, but Apple's failure to continue the original Mac Pro tower range for users who just wanted the Mac equivalent of a boring pick-up truck to get the job done.
I don’t see the point in criticizing the iMac Pro with today’s standard.
Except "today's standard" doesn't exist in the Mac world - there's currently nothing you can buy on the desktop between a fully tricked-out regular iMac (which beats the iMac Pro in a sprint but is hot, noisy and lacks the iMP's extra I/O) and a Mac Pro, which was designed with such
extreme expandability in mind that it doesn't begin to offer value-for-money until you spend about $20k on it - the $6k entry-level specs are a complete joke... and the only Apple Silicon available is still primarily designed for the entry-level, ultraportable laptop market. Just because you can edit 8k on it doesn't mean it's a replacement for a Pro machine.
The Xeon-W CPUs in the 2019 Mac Pro would have required a completely new motherboard as the socket was different and enormous (LGA-3467 compared to the LGA-2066 for the iMac Pro). I don’t think the PSU would ever have been able to handle the extra power requirements.
...and
that's part of Apple's problem with desktops - they design the concept so tightly around a particular CPU and GPU that there's no future development path for them without going back to the drawing board. Which, I suspect, is part of the thinking behind Apple Silicon: now Apple can design the CPUs to fit in the boxes rather than vice-versa.
All this handwringing by some are the people who weren’t going to buy them anyways or who are still trying to keep their 2010 iMacs going, but somehow think that they were the target market.
The "handwringing" arises not so much because of the products Apple makes, but the huge yawning price/performance/form factor gaps in their product line. Anybody who, in late 2017, was still trying to keep a 2010 iMac going
should have been the target market for a "Pro desktop" of some variety.
The only alternatives at the time were the already outdated trashcan or the top-end 2017 i7/Radeon pro 580 5k iMac There was a
huge $2000 price jump between that and the iMac Pro.
Apple
decided to freeze the iMac on the old 2012 chassis, when the i7 option, at least, really needed the iMP's improved cooling and, preferably, I/O (maybe that wasn't possible, or maybe an extra pair of TB3 ports could have been squeezed in if they freed up the I/O lines used by the iMac's redundant SATA support and went SSD only). They decided that, if you wanted something a bit better than an iMac, you'd have to make the jump to Xeon/ECC/Vega Pro. They
decided in 2012 not to continue with the original Mac Pro tower concept which is, I suspect, what a lot of the "handwringers" really wanted.
I bought a fairly-high-spec regular iMac in summer 2017 after
seriously considering both the Trashcan or waiting for the iMac Pro (announced, but not available then). In many ways, the
trashcan would have met my needs better - I really wanted a desktop with 2-3 24" displays - but it was already outdated, old CPU/GPU, no TB3 and, of course, Apple had already had that press conference effectively admitting that it was a dead end. The iMac Pro was (a) not yet shipping and (b) would have cost another $2000
for an all-in-one design which was already not quite what I wanted. If it had actually been available I might still have gone for the iMP, but I had reasons for needing to buy just then and the cost didn't inspire me to wait.
Comments like this [My 13in M1 Mac Book edits 8k footage just as well as my $80k Mac Pro] kinda make me wonder.
Yeah, and my $40 Raspberry Pi can crop and sharpen my photos just fine, too...
I think all that says is that a lot of Youtubers got $80k Mac Pros for TEH LULZ because they'd just monetized their 15 minutes of fame and felt like millionaire movie stars. Otherwise, you don't buy a $80k Mac Pro for simply editing self-important home movies (which, obviously,
have to be in 8k) or other jobs where the CPU/GPU/RAM usage meter is hardly going to twitch.
The M1 edits 8k footage mainly because the M1 GPU has hardware to support it (and has super-fast SSD so you probably don't notice that the swap is getting hammered...) - the Mac Pro is for when you need to edit and composite loads of 8k streams, apply lots of complex effects etc, plug in specialised interface/accelerator cards or otherwise do jobs that
actually need 1.5TB of RAM for something other than keeping Chrome tabs open... Oh, yes, and have all that running for hours or days at a time without instability or thermal throttling.
In hindsight, pro users should have just tried to make do with the iMac Pro instead of fighting for a Mac Pro that most couldn’t even afford. The M1 Pro Macs would have been worth the wait.
There probably aren't going to be
M1 Pro Macs. The M1 can't even support more than 2 displays or more than 16GB of RAM. The GPU is maybe as good as a 4-year-old consumer/gaming desktop dGPU - bloody amazing for an ultra-low-power integrated GPU but no match for the sort of GPUs going into Mac Pros. The benchmarks are mostly showing that the M1 gives a 16" MBP or iMac a run for its money on selected tasks - which is impressive but doesn't make them Pro machines.
Apple now need to show that they can scale Apple Silicon so that it convincingly thrashes machines with i9s and half-decent dGPUs. There's a good chance they will succeed, but it's not guaranteed. Even the high-end 13"/rumoured 14" MBPs will need some sort of souped-up M1X, the 16" MBP/regular iMac will need something with more CPU and GPU cores and more TB3 i/o (call it the "M2") - while the Mac Pro replacement... is a challenge.
Will Apple make a whole new "M3" chip to beat the higher-end Xeon W on CPU cores, RAM support and PCIe lanes? That's a very small market from which to claw back the cost of creating a unique CPU chip. Will they stick with on-chip GPUs, move to on-package GPUs, make their own discrete GPU or just support AMD GPUs in PCIe slots? Pretty sure you can't mount 1,5TB of RAM directly on a CPU package... and will all of that mean that they lose the performance advantages of "unified RAM" and short, on-chip or on-package links with RAM, GPU etc? (I have a wild guess that we might see the "M2" supporting multiprocessor configurations that distribute RAM and GPU between multiple systems-on-a-chip).
Also, the target Mac Pro customers will probably be the last group for whom all the apps, plugins, drivers etc. on which their work depends is Apple Silicon native (or at least reliable under Rosetta) so moving them to Apple Silicon is going to be a big ask anyway.
I suspect that the "real pro" Mac (i.e. iMac Pro, Mac Pro) replacements will be the last machines to be replaced, which is a long time for someone who bought an iMac Pro in 2017 to "Make do". I wouldn't be surprised if there's never a direct replacement for the 2019 Mac Pro (I'm calling MPX cards with M1X/M2 'accelerators' as an early Pro product).
instead of fighting for a Mac Pro that most couldn’t even afford.
I'm not sure how many users were "fighting" for a $50,000 Mac Pro. I think they were mostly fighting for a $3000-$10,000 "pick-up truck" but Apple decided to only listen to the few who wanted a Telsa Cybertruck. What Apple produced certainly ticks all the boxes, then adds a few more boxes and ticks those too, but the result is almost a deliberate parody of the 2010 Mac Pro continuation that most users actually wanted. "You asked for a couple of PCIe slots so we gave you
eight and didn't spare the width!!! User-expandable RAM - you must want
1.5TB!!!"
Problem is, using the same chassis as a $50k super-workstation (with quad GPUs, afterburners, 1.tTB ram etc.) to pick up where the Mac Mini leaves off just doesn't work, which is why the $6k price tag on the entry-level Mac Pro is a joke unless you're planning to add another $10k worth of expansion.
I fear the only people left in the mid/high-end Mac market are those who really, really must have MacOS (out of very string preference or because re-tooling/re-training would be prohibitive) - and Apple is milking them for all they're worth. Show the iMac Pro or Mac Pro specs and prices to a Windows or Linux user and they'll just laugh like a drain - not necessarily because they can match that spec at the price, but because they can get the
precise spec
they actually need for half the price, and they don't care if it's not wrapped in a steampunk-styled solid aluminium sculpture.
The "Mini Mac Pro" rumour is interesting - but I suspect it's going to be more of a "square trashcan" concept than a cut-down 2019 Mac Pro. Which
might make more sense in the Apple Silicon world, where apple should be able to ensure a succession of CPU/GPU upgrades.
