Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Together those CPUs are drawing 170w TDP at base frequency, if they both run at 150% of TDP (255w), they are still under the 3990X.

I’m sure that when Dell feels confident that their 2U rack model can handle the TR3-3990X, they’ll offer it in that chassis. Pretty sure that the OP needed something in a timely manner, with decent pricing and Enterprise support.

Why everyone seems to think that AMD and, by adjunct, the 3990X is the answer to everything is beyond me, because it isn’t. The knee jerk reaction of “just put in a 64 core AMD and that will fix everything” is hilarious, because it won’t.

Also, AMD won't always be the leader and they really haven’t been until they released Zen...they have finally gotten their sh*t together and I wish them well, but they will hit a wall at some point...the streak won’t last forever. Their Navi rollout and driver issues are evidence that AMD is making great progress, but is still not executing on all cylinders.

Yes, Intel’s TDP ratings are complete crap based on the referenced article and others. I suppose I could say something like, if I were the CEO, but I can’t. They need to regroup now before 2020 becomes just another lost year, which it feels like they’ve been in a lost year since Broadwell. WAKE UP INTEL!!!
 
Last edited:
Together those CPUs are drawing 170w TDP at base frequency, if they both run at 150% of TDP (255w), they are still under the 3990X.

I’m sure that when Dell feels confident that their 2U rack model can handle the TR3-3990X, they’ll offer it in that chassis. Pretty sure that the OP needed something in a timely manner, with decent pricing and Enterprise support.

Why everyone seems to think that AMD and, by adjunct, the 3990X is the answer to everything is beyond me, because it isn’t. The knee jerk reaction of “just put in a 64 core AMD and that will fix everything” is hilarious, because it won’t.

Also, AMD won't always be the leader and they really haven’t been until they released Zen...they have finally gotten their sh*t together and I wish them well, but they will hit a wall at some point...the streak won’t last forever. Their Navi rollout and driver issues are evidence that AMD is making great progress, but is still not executing on all cylinders.

Yes, Intel’s TDP ratings are complete crap based on the referenced article and others. I suppose I could say something like, if I were the CEO, but I can’t. They need to regroup now before 2020 becomes just another lost year, which it feels like they’ve been in a lost year since Broadwell. WAKE UP INTEL!!!
Well, Dell offers 2x 64-core Epyc in 1U then.
 
That looks great! Can you tell me what drive setup is below the Rackmac Mini? Thanks!

Primary NAS behind the grille is a Synology DS1517+ with 5xSSDs on a 10GbE link (to both network switches, iMP, MBP and the Mac mini in the rack):

Synology DS1517+ BW.jpeg


Above that is a rather basic 2 bay Synology RS217 that provides HDD backups only.
 
  • Like
Reactions: citysnaps
I have 19" rack equipment in my house… in 2 places. Pretty handy if you ask me. My "problem" (don't have a real problem as I'm not who this computer is aimed at) with it isn't that it's rack mount… its the fact that it's an enormous 5U. If it were 3U I could have happily mounted it into the 19" rack mount that's screwed in under my desk.

It didn't even dawn on me that it was 5u's. that's pretty heft for a rack kit. But this is a unique niche product.


Heck, I can run an entire bank's back end off basically 3u right now.
[automerge]1579118558[/automerge]
[automerge]1579028379[/automerge]


Does Mac OS have some network based setup out of the box?
Does the Mac Pro have an IPMI Interface they are hiding from the rest of us?
Does Apple even support PXE with Mac OS?

Yes. you can install Mac OS via network/internet directly from the EUFI bootloader on almost all mac's since 2011 (around then).

No, Mac Pro is not likely to have an IPMI, even in the rack mounted kit. This is not a server. it's a workstation still.

MacOS is irrelevant for PXE. PXE Boot is a BIOS/EFI boot protocol that is engaged prior to OS. Does the EUFI that ships with the Mac Pro have PXE support? I don't know.
[automerge]1579118594[/automerge]
View attachment 888592
It'll fit perfectly and I can FINALLY open more than 3 Chrome tabs!
No hating allowed on the Lack Rack
 
Apple really wants to quench the appetite of admin/IT staff don't they. "Yes, why go for a Dell Poweredge when we can go for a Mac Pro ?"

businesses i think usually go half and half taking into account money more seriously over the 'best of the best'




Single OS, nope.. but multi boot a Mac and ... they would.

business side stuff is always a "fight" btween accounting, and operations.

Operations want the sky. they want everything. Accounting often times is the one who keeps them in check


it's than up to us in IT to take both sides, and solution. that's the real balancing act that has to happen. Cost analysis that outlines the performance levels versus cost.

from an IT management perspective when we are solutioning and especially when we hvae budgetary considerations. We don't always have the leeway to buy the most expensive stuff, even if it's the fastest.

an Example of some of the tradeoffs forced by this balance:

Buying a new server. (all numbers fictional for ease)

$10,000 monster would yeild processing times of 1 minute
a $8,000 server would yield times of 55 seconds

Does the 5 seconds have material costs associated with it? will it hurt revenue by not being fast enough? Will there be significant impact to those 5 seconds?

Because the cost analysis on these servers would yield the $10,000 server is costing you $166.66 / second
the $8000 server costs you $145.45 / second

it's juts all numbers and cents. and I enjoy doing that sort of thing way too much :p
 
Rack mounting Macs is far from unusual even for home offices - mine in a Sonnet RackMac enclosure:

View attachment 888716:

That looks great, but I fail to see how some people having a rack mounted Mac at home makes it far from unusual. It still seems unusual for someone to have a rack mounted computer at home, let alone a rack mounted Mac.
 
They can do it because they design nothing, they just buy parts from others. They do not develop an operating system (macOS) or productivity apps (Keynote, Numbers, Pages), design custom chips (T2), or custom cases.

CPUs, GPUs, motherboards, software, cases etc. - even those not made by Apple - don't just crystallise out of thin air, you know. Someone has to design them, test them, write the drivers, tool up the factory, advertise them, pay the distribution costs... and if you buy parts from others then those costs just get wrapped up in the price you pay. In fact, the smaller you are, the less bargaining power you have with suppliers and the bigger the margin you'll pay.

Want to sell your PCs with an operating system? Pay license fees to Microsoft. Want to bundle a basic productivity suite? Pay for it... or choose free software like Linux/OpenOffice etc. but then face the cost of supporting it yourself.

Note that we're talking about full-sized desktop/tower personal computers here - there are reasons why something like an iPhone, stuffed full of custom-engineered parts, is hard/expensive and dependent on astronomical sales, but designing a PCIe tower to run an existing OS (even in a nice pretty box with a new tidy way to power PCIe cards) ain't exactly the Manhattan project.

Yes, Apple has the extra costs of maintaining MacOS and their "free" productivity apps (...that doesn't include Logic, FCPX etc. which still cost money) but those are all mature products that haven't seen much in the way of huge updates recently (if Apple are short of cash maybe they should stop breaking MacOS and iOS every year...) - and nobody is expecting Macs without some sort of premium over PCs, we just want a mid-range headless Mac to exist.

Apple is never going to release a machine made with someone else’s motherboard or case.

Last time I looked, Apple's motherboards and cases were mostly made by Foxconn (who make lots of PC parts for OEMs), their display panels were made by LG and Samsung, their GPUs were made by AMD, the RAM comes from Micron, Flash from Samsung.... In some cases they haven't even re-branded stuff as Apple, despite it being vital parts of the system - such as the internal HD expansions for the Mac Pro (Promise), the power cables for fitting standard PCIe cards (Belkin), TB display-compatible eGPUs (BlackMagic)....

So, yes, Apple will absolutely use parts from the cheapest source, and only "vertically integrates" where that is actually cheaper.
 
...and huge economies of scale given the resulting ability to buy in quantity and negotiate tough quid-pro-quo deals with suppliers, not to mention the money saved by selling a large portion of your output directly to customers rather than letting Best Buy, Walmart, Amazon et. al. take a slice.

However, the kind of logic you're suggesting is part of the problem: whatever it might have been in the past, in 2020, the Apple retail store network is there for iPhones, Watches and Apple gift cards (with a side-order of general flag-waving for the brand). There's no way that a personal computer range can come close to justifying that sort of infrastructure (that's why you don't see many Dell, HP and Lenovo stores in prime retail locations).

Likewise, R&D - I'm sure that the R&D bill for the Mac Pro looks impressive quoted in isolation - but I'm also sure that it is a tiny fraction of the R&D budget for the next iPhone (Existing tech like T2 and thunderbolt + reference Intel designs + lots of stock ICs versus ultra miniaturisation of every component, new camera technology, new display technology, new battery technology, fully custom system-on-a-chip...)

If Apple are expecting every Mac - especially the higher-end ones - to pull its weight in Apple Stores and subsidise iPhone R&D, then the Mac is pretty much doomed... but that's "unicorn start-up" thinking: a large corporation making everything from watches to TV shows doesn't need all of its divisions to cleve to the same business model - frankly, they're ahead of the game if they're all in the black. ...or do you think Samsung has the same mark-up on their fridges and dishwashers as they do on the Galaxy Fold? Because that's what personal computers are in 2020 - mature, commodity tech, who's boom time has passed, but which people still need and pay good money for.

The Mac business is still solid and profitable: gimping it because all you can see is that the revenue is 1/10 that of the iPhone will kill it... and the thing about the Mac is that it is an ecosystem that relies on sufficient users and diversity of uses to keep software and hardware developers employed. Affordable machines for power users/developers (no "xMac"), for basic office productivity users (no sub-$1000 MBA replacement) etc. are an essential part of the ecosystem, even if they're not the most popular. You can't build an ecosystem with just meerkats and elephants.



First, that's a nonsense price - you don't make an "xMac" by taking the Mac Pro and pulling bits out. The whole problem with the $6k entry level MP is that its already a $50k machine with the good bits missing: you're paying a huge premium to get a machine with a huge PSU, a motherboard with slots for 1.5TB of RAM and enough PCIe bandwidth to run quad GPUs and an afterburner or two - if you don't want to add $24k+ of upgrades then you're paying $6000 for $2000 worth of performance.

...so, yeah, offer me even a $6000 Mac that is even slightly competitive with the sort of usable spec you can get for $5k+ in the PC world - as soon as you ditch the requirement for 8 PCIe slots and 1.5TB of RAM slots - then I'll at least think about it.

...but, honestly, $3k+ would be a more sensible price for a straightforward Core-X PCIe tower, a target Apple have always managed to meet in the past even with Xeon CPUs (PCIe towers built from standard parts should be cheap to develop and make - the $3k Mac Pro trashcan, completely made from custom parts, would have been far more expensive). Even that would carry a substantial premium over the sort of PC you can get for $2k+ and it wouldn't necessarily deter those legendary "true pros" who need the power of the $50k Mac Pro.
You don’t get huge economies of scale when you’re selling a couple hundred thousand units. A couple hundred million, sure. But Mac Pro is a very low volume product.

But the fact that Apple has large overhead expenses is relevant to every product. You can’t expect iPhones and iPads to subsidize Mac and MacOS. Macs do have to carry their weight, and Apple prices them on that basis. That limits their market share, and so be it.

You must understand that Apple’s cost isn’t simply BOM cost. The fact is, Apple makes about 20% net profit. If components cost 30%, there also another 50% of the sales price attributable other costs.

Those costs can’t be ignored. They pay for salaries; yes, the cost of retail stores; advertising; $18 billion a year in R&D; taxes; and many, many other expenses. Apple can’t sell you $1,500 worth of parts for $2,000—they’ll lose $2,000 on every sale.

Perhaps an example would help. Take iphone. An XS Max that sells for $1,249 has a components estimated at about $445. Apple’s profit is about $250. And then there’s another $555 of other expenses. If Apple sold that iPhone for $1,000, they’d just be breaking even. You can’t run a business on zero profit.

That’s the reality. I know you want Apple to be able to sell you a cheaper Mac. They can’t. Their cost structure doesn’t allow it.

Apple doesn’t play in the cut-throat, low-margin hardware market. You might as well ask why they don’t make $200 iPhones to compete with cheap Android phones. Or $500 laptops like the disposable crap you see at Best Buy. Nope, the Apple versions cost twice that. And the $2,500 tower you want is $5,000. For the reasons stated above.

We’ve gone around and around on this before theluggage. There’s really no point in doing it again, so I’ll bow out if you don’t mind. I can explain it to you, but I can’t understand it for you 🤷‍♂️
 
  • Like
Reactions: Zdigital2015
Apple IS aware that server racks are generally locked up out of sight, in an already noisy environment (e.g., A/C, in-row coolers, etc.) and thus the fancy case laser etched out of billet aluminium or whatever is entirely un-necessary and irrelevant?

Guaranteed they won't be using these in their own datacentres!
[automerge]1579130396[/automerge]
But Mac Pro is a very low volume product

Yeah, because Apple seem determined not to sell a Pro machine that actually makes sense if you value function vs. cost over form.

A shame.

It's not like they don't know how to build a server platform:

 
Last edited:
Together those CPUs are drawing 170w TDP at base frequency, if they both run at 150% of TDP (255w), they are still under the 3990X.

You might want to see just how far intel are pushing power beyond rated TDP in 2019/2020.... only 150% isn't even close in a lot of instances.

Also, 2 CPUs at 255w = more than the 280 watt TDP of a single 3990X with more cores.
 
Apple IS aware that server racks are generally locked up out of sight, in an already noisy environment (e.g., A/C, in-row coolers, etc.) and thus the fancy case laser etched out of billet aluminium or whatever is entirely un-necessary and irrelevant?

Guaranteed they won't be using these in their own datacentres!
[automerge]1579130396[/automerge]


Yeah, because Apple seem determined not to sell a Pro machine that actually makes sense if you value function vs. cost over form.

A shame.

It's not like they don't know how to build a server platform:

This is not a server nor a server platform.

it is nothing more than the Mac Pro in a rack mount kit for those who wish to rack their computer. For some professional workflows, who have numerous other pieces of rackable equipment in small space, it's an efficient use of space.

I've seen similar setups for audio and video production focused people.

Honestly though. If you are looking to run back end server stuff, the Mac Pro is an absolutely terrible value for that purpose as there are actual servers availabe that have server "stuff" that the Mac Pro doesnt have. Such as Redundancy on critical systems, ILO or other integrated management ports.

And most backend stuff will run perfectly fine on a linux server that doesn't need the UI overhead (and cost overhead) of running MacOS.
 
  • Like
Reactions: Alan Wynn
...and huge economies of scale given the resulting ability to buy in quantity and negotiate tough quid-pro-quo deals with suppliers, not to mention the money saved by selling a large portion of your output directly to customers rather than letting Best Buy, Walmart, Amazon et. al. take a slice.

One gets economies of scale based on the quantity of parts for specific projects, not just by being big. Sales through Apple Retail save some margin, but they have to pay for the stores, in store support, etc., meaning it is not that great a savings.

However, the kind of logic you're suggesting is part of the problem: whatever it might have been in the past, in 2020, the Apple retail store network is there for iPhones, Watches and Apple gift cards (with a side-order of general flag-waving for the brand). There's no way that a personal computer range can come close to justifying that sort of infrastructure (that's why you don't see many Dell, HP and Lenovo stores in prime retail locations).

Yes, one would expect that Apple is smart enough to attribute costs to products based on the costs incurred because of the products. Meaning that your xMac is likely to incur more support cost (the questions that hobbyists have are more complex and take more time), and more sales costs (lots more questions vs. “these are the options my kid got from the college, what do they cost?”

Likewise, R&D - I'm sure that the R&D bill for the Mac Pro looks impressive quoted in isolation - but I'm also sure that it is a tiny fraction of the R&D budget for the next iPhone (Existing tech like T2 and thunderbolt + reference Intel designs + lots of stock ICs versus ultra miniaturisation of every component, new camera technology, new display technology, new battery technology, fully custom system-on-a-chip...)

True, but one sells hundreds of millions and one sells hundreds of thousands. Given that they are not on the same architecture (yet) :) all that design cost needs to be absorbed by the product itself and cannot be shared with the high volume mobile products (unlike the AppleTV benefiting from the iPhone).

If Apple are expecting every Mac - especially the higher-end ones - to pull its weight in Apple Stores and subsidise iPhone R&D, then the Mac is pretty much doomed...

Every product should absolutely ”pull its weight” in Apple Retail. What that ”weight” is should be based on costs the product causes the stores to incur (support, sales, inventory, etc.) but should also take into account value to the ecosystem that product adds.

No other product subsidizes iPhone R&D, nor is expected to do so. R&D for the iPhone subsidizes all the other products built with the same parts (AppleTV, HomePod, iPad). However, a product should be expected to cover all its own R&D costs.

but that's "unicorn start-up" thinking: a large corporation making everything from watches to TV shows doesn't need all of its divisions to cleve to the same business model - frankly, they're ahead of the game if they're all in the black

All its divisions need to add to the total ecosystem and unless there is some overwhelming benefit to it need to be profitable including their amortized costs.

or do you think Samsung has the same mark-up on their fridges and dishwashers as they do on the Galaxy Fold?

Given how much they are going to lose on the Fold, I would expect not.

Because that's what personal computers are in 2020 - mature, commodity tech, who's boom time has passed, but which people still need and pay good money for.

Apple serves the bulk of his mature and declining business (desktop/laptop computers) very well with its iMac, iMac Pro, MacBook Air, MacBook Pro, Mac Mini and Mac Pro product lines. The market for a lower-end tower is just too small for Apple to address given their inherent constraints.

The Mac business is still solid and profitable: gimping it because all you can see is that the revenue is 1/10 that of the iPhone will kill it...

Just as forcing it to build products that will necessarily lose money will kill it.

and the thing about the Mac is that it is an ecosystem that relies on sufficient users and diversity of uses to keep software and hardware developers employed.

Which they seem to be able to do with their current market share. Companies like Blackmagic Design, Otoy, Maxxon have ported (or are in the process of porting) their software to the platform. Given the small market share, these companies do it for a reason - the macOS market contains a much higher proportion of high value users.

Affordable machines for power users/developers (no "xMac"),

Sorry, here is the problem with your argument. You have decided that your definition of power users/developers is “those people who want a mid range tower”. Apple serves power users and developers with several different machines: the iMac, iMac Pro, Mac Mini, MacBook Pro and Mac Pro all serve those users with a broad range of price points.

When Apple last made a tower, their own studies, as well as those done by others, showed that fewer than 1% of people ever installed a PCI or PCIe card, and an even smaller number upgraded their graphics card. That was a large part of why the Mac Mini was built in the first place. At its height, that machine sold hundreds of thousands of units - not enough to be sustainable. Even at a million machines and 5% adding cards you are talking about a market for under 50,000 machines, split between your mid-tower and the New Mac Pro. The higher end of that market is much more valuable (not in profit, but in getting marquee software companies to port to the platform).

for basic office productivity users (no sub-$1000 MBA replacement) etc. are an essential part of the ecosystem, even if they're not the most popular. You can't build an ecosystem with just meerkats and elephants.

Again, IBM has now done two separate studies of TCO for Mac systems vs. Windows PCs. The first determined that the TCO for Mac systems was an average of $535 less than the comparable Windows machine (even though the Windows machines were all much cheaper upfront). Their second study showed that not only did the TCO for Mac systems remain lower even after their numbers substantially increased within the company, but those users who received Mac computers had higher job satisfaction and were more likely to remain with the company.

Part of the reason is that many of these “lower priced“ machines get that way by installing bloatware, covering their machines in advertising stickers, shipping trial versions of software, etc., and other things that make users dislike them.

First, that's a nonsense price - you don't make an "xMac" by taking the Mac Pro and pulling bits out. The whole problem with the $6k entry level MP is that its already a $50k machine with the good bits missing: you're paying a huge premium to get a machine with a huge PSU, a motherboard with slots for 1.5TB of RAM and enough PCIe bandwidth to run quad GPUs and an afterburner or two - if you don't want to add $24k+ of upgrades then you're paying $6000 for $2000 worth of performance.

Again, as has been pointed out before, there is a cost to designing and building a totally new machine. While Apple does use contract manufacturers to produce its designs, it does not use Intel’s reference designs and never has. Its cases are also always custom designed. The point that is being made about the high starting price of this machine is simply that most of the additional cost comes from amortizing the cost of the design over the small number that they expect to sell. While there is some cost for the beefier power supply, the extra RAM and PCIe slots and the larger case, they account for tens or hundreds of dollars. There are also some costs for needing a chip that supports the extra PCIe lanes, again, hundreds more. The bulk of the cost is the cost of design spread over a small number of units.

Splitting that small number in half, does not make the cost lower, it raises the cost. While there might be a small decrease in manufacturing and BoM costs for a smaller machine, the design cost per unit goes up.

Unlike those on here that state that their beliefs of the size of this market, Apple had real numbers from the last time they made one of these machine (the previous Mac Pro that some many seem to be dying to buy). Apple did not stop making the machine out of spite, they stopped because it was not profitable. They did not upgrade it because it was not profitable. They justified the design of the trashcan because they knew how few people added expansion cards and/or upgraded their GPUs.

...so, yeah, offer me even a $6000 Mac that is even slightly competitive with the sort of usable spec you can get for $5k+ in the PC world - as soon as you ditch the requirement for 8 PCIe slots and 1.5TB of RAM slots - then I'll at least think about it.

Given that Apple has no shipping products based on AMD CPUs, designing a totally new product based on a new architecture is not cheaper, it is more expensive. The extra PCIe and RAM slots are not the cost drivers you want them to be.

...but, honestly, $3k+ would be a more sensible price for a straightforward Core-X PCIe tower, a target Apple have always managed to meet in the past even with Xeon CPUs (PCIe towers built from standard parts should be cheap to develop and make - the $3k Mac Pro trashcan, completely made from custom parts, would have been far more expensive). Even that would carry a substantial premium over the sort of PC you can get for $2k+ and it wouldn't necessarily deter those legendary "true pros" who need the power of the $50k Mac Pro.

Given that you acknowledge that Apple made a machine at that price point and then stopped, you seem to be arguing one of three things:
  1. Apple just decided one day that they hated the users in this market and actively did not want to serve them.
  2. That in the period since they last made a machine with those specs, that part of the market has grown and desperately wants to buy Macs but Apple does not know this.
  3. Apple is stupid and/or greedy and wants to make a machine that is much more expensive, knowing that it well sell fewer systems because it is targeting stupid/gullible/misguided/trapped people who will buy it.
None of these seems particularly reasonable.

While you may not like the newest iMac, the iMac Pro or the new Mac Mini, Apple has a pretty good idea who is buying them and, based on its customer research, seems to think that the product you want will not be profitable. It is certainly possible that all the people with actual data about costs and sales are wrong and you are right, but I would not want to make that bet, nor does it seem that Apple wants to either.
 
Last edited:
You might want to see just how far intel are pushing power beyond rated TDP in 2019/2020.... only 150% isn't even close in a lot of instances.

Also, 2 CPUs at 255w = more than the 280 watt TDP of a single 3990X with more cores.
Each Xeon is rated at 85w TDP, so 170w TDP for both, but yes, Intel is playing fast and loose with their TDP measurements. Cannot wait to see if the i9-10900K actually draws 300w as rumored. 10-cores and 300w, WTH, Intel?
 
Apple IS aware that server racks are generally locked up out of sight, in an already noisy environment (e.g., A/C, in-row coolers, etc.) and thus the fancy case laser etched out of billet aluminium or whatever is entirely un-necessary and irrelevant?

While it has already been pointed out that this is not designed to be a data center server, almost the same comments you make about this machine with respect to data center use, people make about PC towers in general: “No one cares about how they look. No one will pay extra for a machine with a nicer industrial design.”

Guaranteed they won't be using these in their own datacentres!

Guarantee you are wrong. These machines will be used, at a minimum, for testing certain macOS processes.

Yeah, because Apple seem determined not to sell a Pro machine that actually makes sense if you value function vs. cost over form.

Would you explain what their motivation is for cancelling the pre-trash can Mac Pro is? It seems that if it was as profitable as you and everyone else seem to argue that a new mid-tower would be, Apple must have some reason that they do not want to build it. Is it just spite?

It's not like they don't know how to build a server platform

Funny you should cite the Xserve as an example, as many people at the time of its release made the exact same arguments against it as you make against the new Mac Pro (“Form over function“)
 
Last edited:
So.. this is not meant for a server or anything similar that would be sitting in your rack room. It's a terrible value for that purpose, and terrible overpriced design for unnecessary silent feature in rack room.

Hence, as some says, this is meant for personal use meant for people with several rack devices in their room or office. In that case, Apple needs to offer various TB3 cable options not limited to not-yet-ready to sell 3m version.
 
One gets economies of scale based on the quantity of parts for specific projects, not just by being big.

So you don't think that ordering the components for 50 million iPhones a year from Foxconn would get Apple any leverage when they ask for a quote for 100,000 PCIe motherboards? Okay...

Meaning that your xMac is likely to incur more support cost (the questions that hobbyists have are more complex and take more time),

...and those hobbyists are far more likely to look up or figure out the answers for themselves without bothering support. If they do call support (which is usually the last resort) they're more likely to have sensible questions with all the relevant details to hand and all the stupid possibilities eliminated. If you've ever done time supporting users you'll know that the real hard work is when you're asked "I clicked on the thing and it said error" and it takes 20 delicately-phrased questions to find out if they're even using the software you're supporting.

...that's one of the hidden costs of alienating hobbyists/power users: they're often the front-line support for their family, friends and colleagues, not to mention the ones people ask "what computer should I buy".

all that design cost needs to be absorbed by the product itself and cannot be shared with the high volume mobile products

All what design cost? Newsflash: since 2006 desktop Macs are just generic x86 hardware - apart from the T2 chip (unnecessary in a desktop anyway), which is existing tech developed for the iPhone/iPad. Even if Apple develop their own motherboard, its all based on established reference designs from Intel/AMD. No I'm not saying that the R&D will be zero, but it will be a tiny fraction of something like the iPhone (miniaturised and custom made from the processor up) or even an ultrabook like the Air.

Every product should absolutely ”pull its weight” in Apple Retail.

Yeah, because those $5k XDR displays and $30k Mac Pro configs are just going to be walking out of your local Apple store...

Apple serves the bulk of his mature and declining business (desktop/laptop computers) very well with its iMac

Did you not get the memo? PC sales grew last quarter, while Mac sales fell. Apple have had a 10 year honeymoon while their biggest competitor, MS, worked through the Vista/Windows 8 fiasco and the teething problems with Windows 10. Hence all of that intensive publicity we've seen from Apple targeting Windows 7 users who MS are no longer supporting... er, no, wait, I must have missed it.... but, hey, the most expensive of their laptop models now has a working keyboard and a sensible base SSD/GPU spec, and they've just launched a super-expensive, 5-digit-price machine aimed at a tiny minority who are committed to ProRes video or orchestra-scale Logic Pro, so that will help...

When Apple last made a tower, their own studies, as well as those done by others, showed that fewer than 1% of people ever installed a PCI or PCIe card, and an even smaller number upgraded their graphics card. That was a large part of why the Mac Mini was built in the first place.

Funny, because the Mac Mini was released in 2005 (initially to provide a relatively cheap entry-level Mac), yet the PCIe Mac Pro tower continued until 2013, since when the trashcan has been acknowledged by Apple as a mistake and (criticisms aside) the Mac Pro tower has been re-instated with even more PCIe slots than before (one of the justifications for its high price).

Of course, one might suspect that the real reason for the Mac Pro is because Apple is getting into the TV/Film production industry and it would be embarrassing if they had HP or Dell logos on their end credits... The Mac Pro/XDR display is perfect for the executive who wants to "view rushes" in their corner office...

Again, IBM has now done two separate studies of TCO for Mac systems vs. Windows PCs. The first determined that the TCO for Mac systems was an average of $535 less than the comparable Windows machine (even though the Windows machines were all much cheaper upfront).

Yes, that's all good... except people, even businesses, don't do accounting in terms of TCO. Even for a "sole trader" business, capital equipment (computer hardware) and "cost of doing business" (support, maintenance, consumables) have to be separated for tax purposes. Often (your jurisdiction may vary) the costs are immediately 100% tax deductible while all you can deduct for the capital outlay is annual depreciation - so its actually more efficient to have more of the TCO coming from support/maintenance. That's one reason why businesses often lease equipment rather than owning it - but that just passes the capital vs. costs balancing act upstream to the leasing company.

...any sizeable business will have separate accounts and management for purchasing, support and labour costs, each independently trying to make efficiency savings. All these "the $50k Mac Pro will pay for itself" arguments come from a minority of users who must be in a very privileged position when it comes to making their own budget decisions. Most of the human race have to justify the purchase price of their kit.

Part of the reason is that many of these “lower priced“ machines get that way by installing bloatware, covering their machines in advertising stickers, shipping trial versions of software, etc., and other things that make users dislike them.

I'm not talking about Apple making $300 economy towers thrown together from whatever components are cheap this month (yeah, Dell et. al. only make a profit on those if you buy an extended warranty and a Monster HDMI cable). I'm talking about the sort of PC that you can have custom-built for $3000-$5000 that typically comes with a nice case, brand-name components and a clean Windows install. Looking at one at the moment (from a reputable firm who I've dealt with for years): £5500, so about the same as the base Mac Pro, but that's with a 28 core Xeon W-3175X processor, twice as much SSD, twice as much RAM and a comparable, but newer, WX 7100 GPU... (and that includes the £220 Windows 10 license, so yes those poor programmers at Microsoft get paid) all in a premium Fractal Design (they're good) case.

So, no, it doesn't support 1.5TB of RAM (which would also need a M-series CPU costing 2-3x as much) or 8x8-lane PCIe slots - but that's the point: the MP makes you pay for those whether you need them or not. It's not that the Mac Pro is over priced its that it is over specified in a way that means the $6k option gets you 1/6 of a $36,000 computer rather than 100% of a $6000 computer.

...and, of course, the same site will sell me a Core, Core X or Ryzen system for substantially less if I don't feel the need for the magic Xeon woo-woo.

Thing is, you can waffle on about R&D, the cost of the retail chain and the need to build giant flying-saucer-shaped buildings with designer door knobs in California as much as you like, but its all excuses. If Apple can't make machines at credible price/spec points c.f. PCs (which doesn't just mean 'cheaper') then that's Apple's problem to fix, and anybody who can leave Mac will.

Again, as has been pointed out before, there is a cost to designing and building a totally new machine.

...and, as has also been pointed out before, a Xeon, AMD or Core X tower PC is not a totally new machine, and while the potential sales are a fraction of something like the iPhone, so are the development costs. Unless you insist on designing a bit of modern art with nano-machined dust intakes and re-invented $400 wheels.

People keep banging on about how only a minority of people want desktop/modular macs... then defending the Mac Pro which is aimed at a tiny minority of that minority...

Would you explain what their motivation is for cancelling the pre-trash can Mac Pro is?

They thought the Trashcan was a good idea.
4 years later they had a press conference in which they were forced to admit that it wasn't.

So they not only un-canceled the Mac Pro tower, but replaced it with a parody of the old "cheese grater" and turned the expandability up to 11, hence pricing out many potential buyers and limiting the sales to 1% of the (according to you/them) 1% of users who actually wanted even 1 PCIe slot.

(or perhaps it was just worth it to make sure that Toy Story 5 still has an Apple logo in the end credits).
 
This is not a server nor a server platform.

it is nothing more than the Mac Pro in a rack mount kit for those who wish to rack their computer. For some professional workflows, who have numerous other pieces of rackable equipment in small space, it's an efficient use of space.

I've seen similar setups for audio and video production focused people.

So basically Apple created another SKU for a niche use case for a niche video editing machine 7 years late on the market...
 
So basically Apple created another SKU for a niche use case for a niche video editing machine 7 years late on the market...
more or less. yes.

it's extremely niche. but there are demands for it.


they must have figured the margins would be good enough to warrant it.

because we know there's demand for a standard desktop Mac. They just probably don't think the margins are enough for them on that product. Which would likely outsell this one.
 
Of course people vastly prefer laptops when Apple does not offer a proper consumer desktop.

But Apple does not offer a proper laptop anymore either, so some just dumped the platform.

And the OS has become more idiosyncratic anyway.
Why do people use the word “proper” when they actually mean “my preferred” 🙄

iMac, iMac Pro, Mac mini and Mac Pro are all “proper” desktops. If none of those four work for you, you should buy a machine that does.

MBP and MacBook Air are “proper” laptops, but if none of the four available models works for your purposes, you should buy something that does.

Seems simple enough 🤷‍♂️
 
  • Like
Reactions: Zdigital2015
Why do people use the word “proper” when they actually mean “my preferred” 🙄

iMac, iMac Pro, Mac mini and Mac Pro are all “proper” desktops. If none of those four work for you, you should buy a machine that does.

MBP and MacBook Air are “proper” laptops, but if none of the four available models works for your purposes, you should buy something that does.

Seems simple enough 🤷‍♂️

Having spent quality time with an SGI Indy, a Sparcstation 20 and a NeXTstation in a previous life, my “preferred” desktop would be a Xeon W-22xx-based slab with a PCIe x16 on one side (horizontal GPU), two PCIe (x16, x8) on the other side, 4 or 8-DIMM slots, 4 TB 3 ports, all the other ports and such and an Apple branded 5K sitting up top in Space Grey. I know Apple is never going to give me that. I have made peace with it, but for me that would heaven, proper!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.