MP 7,1 Waiting for Mac Pro 7,1 A1991 (no more)
- Thread starter Mago
- Start date
-
- Tags
- macpro 2016
- Sort by reaction score
I think they saw the writing on the wall. I don't have their remarks on hand, but when they said something about the new Pro, they implied that they'd received a lot of heat from professionals and that they were being dumped for HP's high end workstations made for film and photo editing. I think we're going to see a slick, mini-tower powder coated in space gray with all the goodies on the back, and easily swappable hardware. HP copied what Apple had done with the older Pros and improved on it, and I suspect Apple will go back to that but make something better than HP has done using their original idea.
The reality of the situation is that the Pro sales make up a small amount of their total computer sales, but it's also necessary to keep the pros happy because they drive sales, too. People looking up at pro editors or whatever, or wanting the best, will likely get a Mac Pro over a MBP if their budget allows or they'll get it through the Apple used store with a warranty.
A Pro might get outdated for video editing, but it should be good for photo editing for at least another 5-6 years.
mattspace
macrumors 68040
Realistically, video and photos are more or less solved problems, even with the current hardware - what current Apple hardware doesn't provide a good solution for, and what eGPU isn't going to be any more than a secondrate solution for, is VR.
There's no getting away from the fact that a VR station, first and foremost, needs to be built around user-upgradable GPUs. If your workstation isn't able to be as good as, or better than, and remain as good as of better than any "gaming pc" that costs a quarter as much to set up, and 1/8th as much annually to keep at the cutting edge, then your workstation isn't up to the task - it's like being an iOS developer, with nothing faster or newer than an iPhone SE, or an iPad Air.
I suppose. The way I see it is that slower, or rather hardware that becomes outdated increases PP times for professionals. As video quality standards increase, the time gap increases. And time is money. You lost me on the VR bit. I don't keep up or read about VR. I think we're a long ways from VR being more accessible to the masses without the initial huge upfront costs. EGPU is promising but it adds slowdown than an integrated unit.
Anyway, building your own rig right now isn't feasible due to the high price of GPUs and DDR4 due to miners. For Apple to truly succeed with the new Pro, they need to step away from proprietary internal connectors or setups that limit what kind of NVME card you can use or whatever that may raise issue with those trying to upgrade in the future.
fuchsdh
macrumors 68020
1080p video is certainly solved, but dealing with losses 4K codecs and the like can always use more muscle, especially when previewing effects. Not to mention there's 6K and 8K DIs now. And then there's >1080p compositing and rendering. It's not just VR that requires better hardware.
Leakers on podcasts are full of ****.
There's also the matter of making 4K media more accessible than it already is. Quality 4K televisions have now just become more affordable, especially if you're someone aiming for more than 60+. They were terribly expensive a mere 3 years ago, and just about useless because of the lack of media.
mattspace
macrumors 68040
Where VR is at right now, is turnkey & a lot closer to production-ready tools than most people believe. There's already people working professionally with this stuff.
In terms of cost, here at least you can still get a 1080ti for less than $AU1100-1200. To put that in perspective, an iMac Pro starts at AU$8200 for the Vega 64 version. In further perspective - the top-spec i7 1080ti water-cooled rigs I was looking at a couple of months back came out to around AU$5000 as a turnkey system - computer, monitor, mouse, keyboard, GPU and complete Vive kit including the harness upgrade.
So your annual cost for staying cutting edge in slotbox-land is around $1200/year, including a CPU & motherboard update every few years. In iMac cul-de-sac it's $8200/year (IF Apple and AMD can even put out a new cutting edge every 12 months).
The economics just don't work for anything with non-replacable GPUs, unless you're talking about mobile VR on laptops, and AMD has nothing to offer there, while Nvidia has mobile 1080s on reasonably cheap machines.
Sure, but the iMac Pro seems to (allegedly) be able to handle 4k & 8k editing right now, so in terms of what needs to be in place, a lot of that is just scaling up (and eGPU seems pretty good for rendering tasks), or a bit of wait time while working - but you can still at least play it back at full speed once it's rendered. With VR we're in a situation where if you're not at the cutting edge of performance on the content creation machine, you literally can't experience the results your users / customers will see.
[doublepost=1516775347][/doublepost]
Heh, entirely possible - BUT the fact remains that the FirePro DXXX were just Radeons with extra non-pro memory, they weren't actual FirePros, and AMD dumped the FirePro brand soon after the nMP came out - whether they had already decided to kill the brand, so didn't care what Apple did to it, or whether the nMP stank (both in terms of melting and in undermining the value proposition of the retail cards) killed the brand for them, is open to debate
How is video a solved problem? That industry is constantly evolving at a faster pace than Apple would like to admit.
Look at BlackMagic for instance, their top of the line hardware is in the form of PCI cards for 12G-SDI 8K capture. Their TB3 deck is stuck at 6G-SDI doing only 4K.
Look at BlackMagic for instance, their top of the line hardware is in the form of PCI cards for 12G-SDI 8K capture. Their TB3 deck is stuck at 6G-SDI doing only 4K.
mattspace
macrumors 68040
It's a solved problem in that even if you can't do something in real-time, there's still "make a cup of coffee while it renders". I agree Apple isn't serving that industry well - I've pointed in the past (possibly in this thread, cause EVERYTHING has eventually been or will be said in this thread) to that 8k card specifically as a fundamental problem with thunderbolt as a strategy, and why Video would be better served with a non-future-fragile, generic slotbox, than a future-fragile, targeted appliance.
But video and photos are still a matter of "it will take longer, but you can still get the results". For VR, which is a real-time, and real-time ONLY medium, you can't let it process a bit longer and get there eventually. It's either cutting edge performance, or trying to imagine what cutting edge will look like, based on the performance you have.
What you said is correct on technical merits, but in practice it is not exactly the case. Photography and 2D desktop publishing for instance, are both "solved problems" if we stick to your perspective in that they can be processed delayed, not necessarily real time. But fact of the matter is that if technology is available then you can bet industries will think of way to utilize the extra horsepower to aid efficiency, and sometimes enable previously unaccessible approaches. Otherwise you wouldn't see the constant complain about Adobe's lack of optimization on HiDPI PCs, Lightroom slowing down while working on large MP RAWs etc. And then there is a book designer in the iMac subforum asking if the iMac Pro will speed his high resolution art book projects since the regular iMac is too slow for him. For these scenarios, having a computer powerful enough or not makes a realistic difference.
What I mean is that, to call something a solved problem simply due to it being fast enough for a certain threshold of task is quite an over-generalized statement. But I guess Apple must have similar lines of thoughts when they determined the MBP's and tcMP's specs.
mattspace
macrumors 68040
Sure, most work processes will expand to fill the available resources, and everyone would like everything to be as fast as possible. But print isn't growing in quality and necessary image size every year - we still use 300dpi @ 1:1 size for 150lpi screen print for most magazine quality printing. The only way I can imagine any print book being slow on modern hardware, is if you're not pre-scaling the original images for the target print quality (learned that the hard way crashing a printer's RIP server in the early 90s by including 600dpi source files). As for Lightroom, that's got problems which probably won't be solved by throwing more computing resources at it.
But yes, Apple's philosophy is "it's good enough today, tomorrow, you buy a new one that will be good enough tomorrow", the philosophy embodied best by the iPad. We can hope there's at least some political struggle over this within the company, because at least with VR, GPUs are as I've mentioned, pretty similar to iOS devices in terms of necessary refresh rates, and suffer an outsized penalty with an eGPU solution.
What I find to be assuring, or at least less troubling, is that during the round table I think Craig did mention VR/AR being an anchor point of performance that the MP shall aim at, so even if the machine is not as expandable as the cheese grater, at least we are getting an envelope that encapsulates other disciplines / industries that supposedly are less demanding. If Apple were to fallback into a singularity approach like the trashcan again, at the very least they have to set the bar higher this time.
mattspace
macrumors 68040
I'm not convinced that Apple's high level executives are sufficiently connected to the rest of the human race to know what non-bajillioaire-version humans want, or like. When the most expensive workstation in their product range is an un-noticibly small impulse purchase, in relative terms, by the people making decisions, those decisions can get weird.
William Gibson said:
singhs.apps
macrumors 6502a
Guys, most of you working in media based workflows know that it’s easier to downscale than to upscale ( AI might solve that issue though ... to an extent )
2d isn’t just limited to video. I use many 2d applications for concept work and while the performance is good in the latest version of Adobe CC, you just can’t get lag free brush tracking in 8k+ Images in photoshop. Add a few adjustment filters and performance slows down considerably.
Also simulating natural media paints at hi res dimensions isn’t a solved issue. There is a lot of room for growth there.
But the benefit of targeting hi end results via hardware is that it can solve many related issues in other areas. If a GPU can push 8k per channel in stereo at 60fps minimum for VR, it can certainly benefit 8k for video production.
Side note: I wonder if Apple’s southern.. I mean Hollywood/Entertainment aspirations has something to do with the revival of the Mac Pro lineup ?
Intel has setup shop as has AMD in Hollywood.
2d isn’t just limited to video. I use many 2d applications for concept work and while the performance is good in the latest version of Adobe CC, you just can’t get lag free brush tracking in 8k+ Images in photoshop. Add a few adjustment filters and performance slows down considerably.
Also simulating natural media paints at hi res dimensions isn’t a solved issue. There is a lot of room for growth there.
But the benefit of targeting hi end results via hardware is that it can solve many related issues in other areas. If a GPU can push 8k per channel in stereo at 60fps minimum for VR, it can certainly benefit 8k for video production.
Side note: I wonder if Apple’s southern.. I mean Hollywood/Entertainment aspirations has something to do with the revival of the Mac Pro lineup ?
Intel has setup shop as has AMD in Hollywood.
And this is exactly why a scalable machine in the form of a workstation tower has been the norm for decades. Having minimum amount of artificial barriers that keep the machine from being utilized in various scenarios that the original maker simply doesn't need to worry about. Just raise the ceiling.
Sometimes I think the success of iPod and then iPhone gave Apple a pretty wrong impression that they must know best for most. At least in creative industries it doesn't work like that, everyone is an adopter.
the tcMP was genial if not so thermally constrained, they targeted 450W system (reasonable 5 yr ago when VR/AR/AI just begin to gain momentum), but then developer required systems for the new nVidia or AMD GPUs with more than 300W on a single card the tcMP needed either to grow in size to increase it TDP or a major redesign.
forget std PCIe cards, APPLE wont allow PCIe cards again unless you use a TB3 adapter, I consider the PCIe card slot has close to 0 chance to be part of the mMP (indeed Apple names Modular the Mac mini too, so dont get scared if the new mMP its just an bigger trashcan)
Define Stupid (Apple does is to collect money from users).
LOL... a Mac? LMAO... Whatever will look the mMP dont expect see rise the Cheese Grater or PC-LookAlike Mac Pro.
I consider Apple will offer either an revised tcMP with bigger thermal core and all the stuff required to handle dual 350W GPU plus the most powerful CPU (single socket) and dual SSD, and maybe 256GB ram ( 128gb as the iMac pro more likely, at least as BTO), expect an mostly sealed system or a system that requires authorized service (and Apple sourced parts) for upgrades/updates.
an important thing is about the GPU bus, Apple need to adopt some architecture compatible with PCIe5 from the begining, and likely proprietary to discourage major DIY upgrades (but available thru authorized partners).
I see more something Modular only on the Assembly line, maybe allowing only RAM and Storage DIY upgrades (as long you find the right parts), and allowing GPU upgrades from Apple authorized partners.
Last edited:
Naimfan
Suspended
That would count as stupid and the beginning of the end of the Mac.
Because it would not come close to addressing the issues the people Apple apparently listened to. A closed system is NOT what they need.
What you described has even less of a niche case than the iMac Pro. At least the iMac Pro has a foot to stand on being an AIO, so cramming everything inside the chassis makes marginal sense for some use cases.
The biggest problem with the tcMP isn't what Apple said, thermo issues or difficulty to upgrade is secondary IMO. The major deterrent is how it doesn't fit in people's workflows no matter how you config it.
I am almost convinced that if we aren't getting a Cheese Greater 2.0, then the mMP has to be a cyber prototype tier product a la the Razer Project Christine. Otherwise why would Apple even bother. The kind of bad press from pros will be much worse than them doing nothing in the last few years.
The biggest problem with the tcMP isn't what Apple said, thermo issues or difficulty to upgrade is secondary IMO. The major deterrent is how it doesn't fit in people's workflows no matter how you config it.
I am almost convinced that if we aren't getting a Cheese Greater 2.0, then the mMP has to be a cyber prototype tier product a la the Razer Project Christine. Otherwise why would Apple even bother. The kind of bad press from pros will be much worse than them doing nothing in the last few years.
Biped
macrumors regular
Perhaps you miss my point.
If I'm a developer, and my product is scientific modelling of protein enzyme digestion ( or make something else up ),
that has certain hardware requirements ( whether is be GPU specific libraries, or PCIE for custom machine control, etc ),
that might have certain software requirements ( relatively current version of gcc, C# RE, JS as well as Klingon, etc ),
and Apple does not have a product that meets those requirements.
Why would I bother going through adding osx as a target in my build process, invest the time in osx specific documentation, divert QA resources to research accommodate osx specific issues, and retrain support staff to support osx.
Even if Apple did have a product that for a couple years did fit the requirements. I would assume that none of my clients would be stupid enough to invest, unless there was some sort of commitment from Apple to stay in this space. Because there are going to be costs for them to migrate to and from Apple products as well.
Flint Ironstag
macrumors 65816
Try thinking of it this way: It provides sufficient boost for practically any TB3 enabled ultrabook to at least experience VR.
let me correct: " I think Apple Listened to"
I have N things against the SEALED iMac pro, but Apple practices has little to do with Machine Niches, actually DIY upgadable machines are not an Apple Niche, even actually most corporate (top500) do not upgrade old workstations, just move to less demanding duty and progressively replace it with complete optimal units, there Its more economical than just upgrade and upgrade unti it deads (upgrades implies downtime sometimes more expensive than the savings).
What you mean by workflow? having integrated DVD, tons of Spinner HDD into a single chassis? LMAO.
Pro users what needs is fast storage, lots of ram and CPU/GPU cores, access to internal components are not part of any PRO workflow.
-hh
macrumors 68030
This is a problem with losing touch with the mortal customers.
And FWIW, I was doing some calculator noodling last week, thinking about how many Mac Pros could be bought for $500B (repatriation money) and then sold for $1000 each....
singhs.apps
macrumors 6502a
Ability to insert a new device of your choice in the middle of a workstation refresh is very much part of a Pro workflow... esp parts that don’t come as a BTO or offered by your vendor.
Why do you think PCIe exists ?
Sorry but closed systems have historically lagged behind open systems ( or systems that are broad enough to be flexible) If not macs and iOS devices would dominate. They haven’t.
Apple might be lining up it’s pockets with its strategy of locking things down... but it’s a harder sell in professional fields than consumer ( read iOS ) devices. If not then your thread wouldn’t have 300+ pages of comments.
AidenShaw
macrumors P6
This is completely missing one of the main points of having a flexible, PCIe-based standard component system.
I agree completely that upgrading internal components isn't that common in large deployments.
However, *choosing* the initial configuration from a wide array of standard options is quite common. Compare the BTO configs for the trash can and any HP Z-series. It's comparing "any color as long as it's black" to the Pantone palette.
And I think that VR/AR/AI/ML will challenge your position. You can't be competitive with a two year old GPU. (Or in the case of the trash can, a six year old GPU.)
I bought my big AI/ML servers with Maxwell GPUs. Those have all been swapped for Pascal GPUs. I'm getting quotes to upgrade to Volta GPUs.
(...and we're Fortune500...)
mattspace
macrumors 68040
An off the shelf i5/i7 with a small ssd and spinners for bulk storage, with a 1080ti is more "Pro" for VR than the most expensive iMac Pro. VR Requires big GPU, and upgrades as often as GPU makers (let's be realistic here, as often as Nvidia) put them out. That feature is overwhelmingly what makes a "Pro VR workstation". Without that ability, Apple does not have a "Pro" VR station, pure and simple.
Worse than that, without that ability Apple doesn't even have a VR consumption platform, because that same i5/i7 which is a better Pro VR machine than their most expensive product, also happens to be cheaper & better for VR than Apple's consumer hardware.
macOS has no part to play in the VR toolchain - your apps come from Steam, and no part of the app experience is furnished by the host OS. So, how exactly is Apple going to appeal to VR creators and app makers, as they seem to have claimed they intend, by offering a machine that has only downsides? Apple talked about how easy it was to port Steam VR apps to macOS - well as of last weekend there's a grand total of 1 VR utility app for macOS on steam, vs 149 for Windows.
In the past people went on about how gaming performance doesn't matter, and how Apple optimised for "pro" apps, well gaming performance IS VR performance. Optimising for gaming performance IS optimising for pro VR apps.
Last edited: