What the MacPro should look like - https://www.servethehome.com/superm...l-xeon-phi-x200-developer-workstation-review/
Two KNL 68-core chips.
Two KNL 68-core chips.
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.
What iMac Pro tells us about MP 7,1
- Thread starter Flint Ironstag
- Start date
- Sort by reaction score
Just looking at the specs of an iMac Pro, "hot and heavy" are the words that come to my mind. Heavy components that produce massive amounts of heat, in a very small space..
It tells me that if someday I buy one it will probably also be made from used parts from a seller on eBay.
Yep, if you are waiting on Apple to give you the option to run Nvidia on the upcoming Mac Pro, I think you're gonna be disappointed. Yes, you might be able to hack it together on an external GPU but it's likely to be unsupported by Apple.
If the iMac Pro is any indication, we're stuck with AMD and custom GPU's.
I don't know that they'll lock out Nvidia for eGPU, would seem to sort of directly contradict the point. But, it does seem AMD GPUs are the way forward.
And I think this is a good thing. It seems to me that AMD is always finding a way to get the most computer power/dollar into its GPUs, and at the same time OpenCL has really not taken off the way it needs to in order to really be any sort of alternative to CUDA. Apple seems to be doing all the heavy lifting on that side of things with Core ML and Metal 2...could be a deadly effective combo with AMD GPUs that are targeted towards not only single-precision compute power but double and half precision as well.
And I think this is a good thing. It seems to me that AMD is always finding a way to get the most computer power/dollar into its GPUs, and at the same time OpenCL has really not taken off the way it needs to in order to really be any sort of alternative to CUDA. Apple seems to be doing all the heavy lifting on that side of things with Core ML and Metal 2...could be a deadly effective combo with AMD GPUs that are targeted towards not only single-precision compute power but double and half precision as well.
So, there's no real alternative to CUDA - but it's good that Apple is going with the low-priced brand?
And the recent AMD chips adopt the anemic FP64 performance of the gaming Nvidia cards.
This was the pipeline Tim has long referenced.
However, does anyone get the feeling this was a pet project or branch off of the iMac design curve that got elevated to the fast-track? Maybe that's contradictory to my opening line but I"m sure someone had played with the idea of a iMac Pro in the past but it wasn't moved on.
My feeling is the Mac Pro 2018 will not have the cost of the screen so the box will be cheaper for that, you can spend that on some extra spec. and you'll be able to upgrade over time which is valuable too many. So they are not totally comparable but if the range of choice remains users are the winners I feel. This act as a bridge between worlds so the distance between steps lessens and that has to be welcomed.
In terms of thermals. The spatial area of the iMac is probably rather suitable forma factor to deal with thermals as it's a large area due to it matching the screen and they can add depth without it being every noticed or appearing on your desktop (deeper case) so to work with and airflow may not be as much as a problem as people perceive. Also noise I would imagine will not spill out as much from behind a large wide plate as much as it might from a box. So it might not be an issue in reality, assuming they use quiet fans etc. etc.
Feeling very positive about this overall.
However, does anyone get the feeling this was a pet project or branch off of the iMac design curve that got elevated to the fast-track? Maybe that's contradictory to my opening line but I"m sure someone had played with the idea of a iMac Pro in the past but it wasn't moved on.
My feeling is the Mac Pro 2018 will not have the cost of the screen so the box will be cheaper for that, you can spend that on some extra spec. and you'll be able to upgrade over time which is valuable too many. So they are not totally comparable but if the range of choice remains users are the winners I feel. This act as a bridge between worlds so the distance between steps lessens and that has to be welcomed.
In terms of thermals. The spatial area of the iMac is probably rather suitable forma factor to deal with thermals as it's a large area due to it matching the screen and they can add depth without it being every noticed or appearing on your desktop (deeper case) so to work with and airflow may not be as much as a problem as people perceive. Also noise I would imagine will not spill out as much from behind a large wide plate as much as it might from a box. So it might not be an issue in reality, assuming they use quiet fans etc. etc.
Feeling very positive about this overall.
Last edited:
I have been told the normal iMacs are not quiet machines under load, so I am rather concerned about what the iMac Pro will sound like.
With the iMacPro Apple should find a decent sized market of folks who will pay that kind of money for a well designed machine that can fly through their typical day. It's my contention that there are plenty of creatives for whom the iMacPro is plenty powerful and will neatly replace their cMP/nMP/aging iMac.
The price is gobsmacking, but Apple has never aimed at the budget buyer. I'd also note that replacement cycles for computer hardware have increased dramatically over the last decade - so to keep a similar revenue stream over time, the vendors need to make more per purchase. Making $800 in profit every 3 years = $1,600 profit every 6 years.
In any case, I do see some signals about the 7,1 mMP embedded in the choices Apple made for the iMacPro.
1) We aren't the only ones who can imagine a scenario where the mMP ship date slips into 2019. By stuffing the iMacPro with higher spec components and getting it out the door by year's end, Apple will have a modern machine to sell while the mMP is in R&D. I think its important to keep enough active high end market that vendors of key software are more likely to continue to develop/port for OS X.
2) If Apple is serious about "modular", then the base mMP model might be as low as $2,999 - but that might only include a lcc CPU/lower spec 4GB GPU and not much else. I think most people who are looking for a beast to handle challenging workloads will end up spending several times that. I do like the idea of being able to pick and choose which components best fit particular requirements. Some tasks will use 512GB of RAM if available, others never access more than 8GB regardless. Some want lots of cores, others need better single core clocks. Etc.
3) I expect the ASP (average selling price) of 7,1 workstations to be pushing $10,000, fully loaded pushing $15,000. If the performance is there, the market will bear it. As many have noted on this forum, a professional tool just needs to earn more than it costs to be viable.
The price is gobsmacking, but Apple has never aimed at the budget buyer. I'd also note that replacement cycles for computer hardware have increased dramatically over the last decade - so to keep a similar revenue stream over time, the vendors need to make more per purchase. Making $800 in profit every 3 years = $1,600 profit every 6 years.
In any case, I do see some signals about the 7,1 mMP embedded in the choices Apple made for the iMacPro.
1) We aren't the only ones who can imagine a scenario where the mMP ship date slips into 2019. By stuffing the iMacPro with higher spec components and getting it out the door by year's end, Apple will have a modern machine to sell while the mMP is in R&D. I think its important to keep enough active high end market that vendors of key software are more likely to continue to develop/port for OS X.
2) If Apple is serious about "modular", then the base mMP model might be as low as $2,999 - but that might only include a lcc CPU/lower spec 4GB GPU and not much else. I think most people who are looking for a beast to handle challenging workloads will end up spending several times that. I do like the idea of being able to pick and choose which components best fit particular requirements. Some tasks will use 512GB of RAM if available, others never access more than 8GB regardless. Some want lots of cores, others need better single core clocks. Etc.
3) I expect the ASP (average selling price) of 7,1 workstations to be pushing $10,000, fully loaded pushing $15,000. If the performance is there, the market will bear it. As many have noted on this forum, a professional tool just needs to earn more than it costs to be viable.
Source? My company is following the same rule that it followed in 2001 - laptops and desktops are bought with three year all-inclusive warranties, subjected to capital expenditure depreciation for three years, and replaced when they are amortized and the warranty expires (at the same time).
For professional use, I doubt that much has changed in this regard.
Maybe the amateurs and pro-summers are keeping systems longer, but the game hasn't changed for the people using the tax codes to their advantage.
LOL nonsense.
If a comparable Z-series is $6K, and loaded $8K - Apple will be handing HP a gift.
If you can bill $15,001 bucks for a job using an Apple that costs $15K, vs billing $15,001 bucks for a job using an $8K Z-series - which will you do?
$1 profit, or $7K profit?
1) Not arguing that many businesses will remain on 3 year replacement cycles, just that a forum filled with people milking extra years out of cMPs might take into account projected longevity in determining what they can ROI. FWIW, most of the companies I interact with that follow the 3 year cycle don't allow Macs anyway, but I digress.
2) Many people who buy Macs believe that it enhances the productivity of the user vs the same user on a Windows machine. If you don't buy into that idea, then it's a moot point. As a 30 year+ Mac user, I have consistently been able to push projects through faster with less headaches vs similar users on Windows machines. (I also have a lot of ProRes contracted deliverables). So, for me at least, the Apple tax is a bargain. Not claiming that's true for all.
2) Many people who buy Macs believe that it enhances the productivity of the user vs the same user on a Windows machine. If you don't buy into that idea, then it's a moot point. As a 30 year+ Mac user, I have consistently been able to push projects through faster with less headaches vs similar users on Windows machines. (I also have a lot of ProRes contracted deliverables). So, for me at least, the Apple tax is a bargain. Not claiming that's true for all.
We're all just speculating, but my guess is that a base 7,1 would come in configured about like a base iMac Pro, spec-wise (but with expand-ability) and cost-wise (perhaps a little less on cost given that it won't have a mandatory screen, unless they prefer pushing the iMacs*).
Personally, I'd be satisfied with a 7,1 configured about like that, though I'd inevitably upgrade some components per their availability. But I do demand that it be in a separate box that can be placed off of my desk (I have my 3,1 in a different room, even).
I've gotten fairly used to my linux box (about $1.5k with quite decent specs), though, and I wonder just how badly I'd be tempted by a 7,1 if it has a minimum price of $4-5k before upgrades or adding what is sure to be a delicious and very expensive monitor.
* The sealed-in nature of the iMac Pro may be foreshadowing a higher Mac Pro price, though, since it could drive people to spend more just for the privilege of adding ram, even.
Personally, I'd be satisfied with a 7,1 configured about like that, though I'd inevitably upgrade some components per their availability. But I do demand that it be in a separate box that can be placed off of my desk (I have my 3,1 in a different room, even).
I've gotten fairly used to my linux box (about $1.5k with quite decent specs), though, and I wonder just how badly I'd be tempted by a 7,1 if it has a minimum price of $4-5k before upgrades or adding what is sure to be a delicious and very expensive monitor.
* The sealed-in nature of the iMac Pro may be foreshadowing a higher Mac Pro price, though, since it could drive people to spend more just for the privilege of adding ram, even.
It never had much anyway, at least not in fora like this one. Everyone had their own definition and used it to push their own opinion of what Apple should make and of why the current lineup was not "pro".
This is the only reason why Apple is still somehow relevant in the video business, but the writing is on the wall
Apple added no depth there at all. The iMac Pro and the iMac 27' have exactly the same dimensions. Height 23.2 inches Width 25.6 inches stand depth 8 inches
https://www.apple.com/imac/specs/
https://www.apple.com/imac-pro/specs/
The enclosure itself is about 2 inches less than the stand depth. (so no more than 6 inches). It is less "deep" than the Mac Pro 2013 ( 6.6 inches ).
They have cut a larger air inlet in the back of the iMac Pro (a slit that goes across almost the entire bottom, back ) and increased the size of the air output vent ( and added a second fan to push air out that expansion). However, if the same same enclosed volume as before. They have nuked the 3.5" HDD drive space and shifted the layout for the bigger cooling subsystem.
The iMac's try to pull heated air down and then out through a 90 degree turn. They've got issues with airflow. The upside is that they heatpipe the major heat sources directly to the exhaust vent, so those two don't radiate much into the case. But there are more than just two heat sources in the enclsoure. Pushed to the upper thermal edge with maximum component usage what they have is not a slam dunk in terms of optimization. It is better than the 'regular' iMac 27" , but we'll see for maximum loading.
Probably about right - my guess would be that they're not going to price it so that a Mac Pro + 5k display costs less than an iMac Pro.
I wonder if the new Mac Pro is going to be as upgradeable as people seem to be hoping. Nothing in the April mea culpa interview promises user upgradeability and they use "upgrade", "update" etc. in the context of Apple keeping the product up-to-date.
...another reason I wouldn't be surprised by a non-user-upgradeable Mac Pro. Personally, I quail at the idea of a $5000-$15000 computer that can't have its RAM, SSD or GPU upgraded mid-life by the user - but then I'd probably be spending my own money and hoping to get something that was likely to last me 5-10 years. However, Apple's target market may well be dominated by 3-year-business-leasers - and there's absolutely no incentive for Apple to let them get away with using cheap third party RAM and SSD from day one.
If it starts off with a 6 core and a less than Vega 52 GPU then why not ? There is zero reason why the Mac Pro needs to start off at the same core count level. Even less reason for the GPU to start off up there. There is still the huge gap of some folks they left behind when the Mac Pro shifted from the $2000 to $3000 entry price range. Pushing that up into the $5+K range is pure death spiral pricing.
Some folks don't need bleeding edge graphics (music production, print graphics, et.c ) . Something like a follow up to Polaris 10 with GDDR6 RAM would probably work for the entry level folks and make a $2,500 price point much more practical. The Mac Pro is suppose to be the answer for those whom the iMac all-in-one form factor has substantive need mismatches with. That is just not necessarily just purely more x86 cores at an even even higher prices.
That can't possibly be "any 5K display". In a year those prices are going to be substantially lower. The LG UltraFine is $1,299. A starting price of a Mac Pro of $3,700 is a $1,200 jump from the zone the MP 2009-2010 model was in. That would be a bonehead move. That is even up about $800 from where the MP 2013 started off at. The jump of the 2013 from 2010 was just around $400. $800 is double that.
The other issue is that version 2 of the iMac Pro may shift back to having a 6 core entry point. ( for example Apple brings ProMotion (120GHz refresh) to the iMac Pro first and needs to throw more money at the base GPUs. The point is that Apple can do something more with the integrated screen to differentiate it from the "race to the bottom" 5K screens that will start popping up in 1-2 years.
Upgrades that don't solve the primary problem aren't going to work. The off-the-shelf video cards don't mesh well with thunderbolt. Since they don't solve the problem, Apple probably won't go there. However, that doesn't mean that Apple has to block off adding more functionality that the Mac Pro doesn't ship with. If they get rid of the "Compute' GPU then that is potentially an empty slot that could be used for something. A few examples
a. A user selected 2nd GPU (which will make the Nvidia zealots happy).
b. a audio DSP card.
c a proprietary a x4 card that some folks can't live without.
d. video capture card for multiple hi res video cameras
That other slot could all be done without getting into a pissing match about evicting Apple's primary GPU solution (that probably won't be replaceable by random off the shelf solution). The Apple primary GPU may be a limited Apple keeping it up to date path. Upgrade and scope expansion aren't the same thing.
I don't think Apple is going to selling what is essentially a 'bare bones' box where the max out the empty slots, bays, and folks fill in what they want. But 1-2 slots covers far more than zero and still puts constraints on power consumption and noise. Similarly more than 1 storage drive can be done without trying to be a all-in-one NAS solution.
Apple's vintage/obsolete policy doesn't run 10 years. I didn't run 10 years 10-15 years ago either. There is a huge gap between 3 year churn and folks squatting on stuff from 7-10 years. No Apple isn't going to enable or promote the latter. They haven't before and they aren't now.
I'm sure Apple will refuse to see it, but the only reason I'm still rolling with 7-10 year old tech is that the 4-year-old tech they offered did not actually meet my needs. If they'd had a mac pro that was of some use to me available in 2013 i'd have been all over it.
Apple should just drop imac pro and give us user upgradable ability mac pro and take our money.
I've got to be honest with you, but opinions like this bemuse me.
Over the last 20+ years from the days of USENET to these forum boards, I have seen Apple enthusiasts beg and lament for a configurable tower that wouldn't charge "pro" prices. These are usually people who want (pro) Macs at commodity or DIY PC prices.
Here's the truth guys. Make an effort to accept and roll with it...it's not going to happen. That's not the way Apple rolls.
It didn't roll that way when Steve was alive, and it certainly will not roll that way with Tim Cook. Whether you like it or not, for better or for worse, Apple with Jobs and now post-Jobs as a profit leader will likely never sell hardware, particularly desktop/workstation hardware, at cost. You are not going to get a tower that you can put your own CPU in. It will likely come with the 8-core Xeon as a base and ramp up from there. Perhaps you go dual core with another module. Perhaps you will be able to run multiple GPUs and use AMD GPUs (NVidias will be perhaps not completely unsupported, but dis-incentivized).
It's not going to be cheap because Apple, especially now, doesn't do cheap. You want cheap, you're going to have to keep Hackintoshing. The high end specced-out Mac Pro will likely be five figures (probably approaching $15,000 or so), and folks that can pay that will for their needs because they probably spend tons of money on speed and storage. The Mac Pro isn't intended, and never has been, for enthusiasts who like having the latest, greatest and fastest Macs for bragging and gaming purposes. A person spending their cash for those purposes are wasting their money, in my opinion.
With all due respect, why keep asking this from Apple when we all know that this isn't what they do?
Perhaps with this new Mac Pro, it will be better than the round can version, but it probably won't quite be the DIY PC either.
Over the last 20+ years from the days of USENET to these forum boards, I have seen Apple enthusiasts beg and lament for a configurable tower that wouldn't charge "pro" prices. These are usually people who want (pro) Macs at commodity or DIY PC prices.
Here's the truth guys. Make an effort to accept and roll with it...it's not going to happen. That's not the way Apple rolls.
It didn't roll that way when Steve was alive, and it certainly will not roll that way with Tim Cook. Whether you like it or not, for better or for worse, Apple with Jobs and now post-Jobs as a profit leader will likely never sell hardware, particularly desktop/workstation hardware, at cost. You are not going to get a tower that you can put your own CPU in. It will likely come with the 8-core Xeon as a base and ramp up from there. Perhaps you go dual core with another module. Perhaps you will be able to run multiple GPUs and use AMD GPUs (NVidias will be perhaps not completely unsupported, but dis-incentivized).
It's not going to be cheap because Apple, especially now, doesn't do cheap. You want cheap, you're going to have to keep Hackintoshing. The high end specced-out Mac Pro will likely be five figures (probably approaching $15,000 or so), and folks that can pay that will for their needs because they probably spend tons of money on speed and storage. The Mac Pro isn't intended, and never has been, for enthusiasts who like having the latest, greatest and fastest Macs for bragging and gaming purposes. A person spending their cash for those purposes are wasting their money, in my opinion.
[doublepost=1497678072][/doublepost]Not if they truly make the MP configurable. Not everyone needs 8+ cores and the best graphics cards money can buy. A LOT of people (like myself) use the tower for expandability (drive bays), user configurability (RAM slots), low noise (have yet to hear a quiet iMac or MBP under heavy load) and greater choice for display options. (not everyone needs 4 & 5K displays).
I didn't consider the nMP because lack of internal expandability, low yet expensive internal drive storage, and having to pay for a 2nd graphics card I'd never use. Love my 2012 cheese grater!
With all due respect, why keep asking this from Apple when we all know that this isn't what they do?
Perhaps with this new Mac Pro, it will be better than the round can version, but it probably won't quite be the DIY PC either.
With all due to respect, why do you assume they wouldnt? Why do you assume as if you work for apple? Because they knew they couldnt upgrade with nmp.
There is a very high likelihood that Apple knows more about the size of the market they skipped by dropping products than the individual customers do. Folks waiting on Apple to do a new XServe or Printer have an even longer "wait' because Apple left those markets.
It is unlikely Apple is going to make a clone of the Mac Pro they made 10 years ago. Four 3.5" SATA drive bays ... probably not. Optical drive bays ... extremely probably not. Some customers may need that, but Apple likely not going to loose any sleep over the size of that submarket.
From the roundtable, 'pow wow' meeting held earlier this year it is apparent that Apple knows there are folks outside the bounds of what the Mac Pro covers. So it isn't "refuse to see". Even if they don't eventually do a complete MacPro 2007-2010 clone it still isn't refuse to see. They could very well see a shallow, stagnant submarket and choose not to participate in it.
Like a lot of folks on this board, I've been biding my time until Apple refreshed the 2013 Mac Pro before replacing my existing Mac Pro. Mine is the 3,1 from mid 2008, which has served -- and continues to serve -- me well, but of course chugs along trying to keep up with advances elsewhere in s/w and h/w. The possibility of an iMac with workstation specs is indeed intriguing. At first I agreed with those who speculated that if an iMac Pro would cost "x," the promised modular Mac Pro would surely cost some multiple of that. But it has since occurred to me that Apple could just as well introduce two lines of professional desktops, with neither being more costly than the other. Some video and photo professionals aren't necessarily interested in expandability, since they're interested in return on investment -- how effectively (quickly) can they satisfy their clients. Assuming they're profitable, they can simply trade in their existing h/w for the next model of iMac Pro and keep on working. Other professionals, research scientists, say, may be much more interested in expandability, since h/w investment costs must be considered in the longer term.