Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple is showing its very weak point with the Mac Pro, they dont undestand/hit its Market, it guru's are lost and blind about.

Assume there is an 2018 Mac Pro, modular(whatever) etc, which it means on of two things:

  1. If the iMac Pro didnt sell as good as they think, so they decided to force inventory cleanup by holding the Mac Pro releasem s they dont care on the Pros but on Cook store manager "vision".
  2. If the iMac Pro sells good, and they bring the mMP a Year Later, means they dont care on those Pro's with Higher requirements, or worst they dont have resources to develop multiple Mac a year.
Common Factor (And Actual Apple's weakness): THEY DON'T CARE ON THEIR PRO'S USER BASE.

What it means on the long? This is the closest equivalent to Ecosystem Cancer, its like killing your father before you learn to walk by yourselves.

Meanwhile, Google is Building Apps and Content on Linux, Windows, and its rumored they will release a custom Desktop Distro based on RedHat or Debian to help solve the mess which means for Linux users to configure/support SOTA Hardware, which it means you buy the Google's blessed Hardware Listed For their Distro and you dont need to worry on configuration, etc, and even get support from Google.
 
No chance IMO. I don't even see how that request makes any sense for a 2019/2020 Apple desktop.

Unless Apple radically changes their approach to SSD pricing in 2019/2020, the need for HDDs in Apple desktops will remain. The iMac 21.5" starts off with no Fusion Drive ( just a plain 1TB ) drive, but to get to Fusion (a small 32GB drive) it is $100 more and to SSD only it is a cap of 256GB (and $200 more). The iMac 27" starts off with a 1TB Fusion drive but to get to 1TB SSD it is $700 more. The iMac Pro charges $800 to jump from 1TB to 2TB of SSD storage ( a 1TB jump).


$100 for a 32GB SSD. To put that in perspective Intel is charging $77 for 32GB Optane drive they released last year. The $100 is closer to the newer and larger ones released this year.
https://www.anandtech.com/show/12512/the-intel-optane-ssd-800p-review

A 1TB TLC 2.5" SATA SSD drive is in the sub $300 zone now. While Apple's PCI-e SSDs have a much faster I/O bandwidth, the $/GB is just way different for more than $700 for a 1TB of capacity.


Apple seems to be conflicted with a "everyone skips into the future on APFS only SSDs" viewpoint on one hand and a "making affordable systems " viewpoint on the other. That's for desktops in general. Either Apple needs to fix their APFS ability to get along with HDD and hybrid systems or need to do a 180 turn on goosing SSD prices artificially high.


In the iMac Pro space, the base price is a price point that more folks will have money to make that trade-off so the $/GB if SSD is more tolerable. The iMac Pro tosses the HDD for more cooling subsystem space which it needs. The targeted users are OK with DAS and NAS solutions for bulk storage.

For a Mac Pro if it is not literally a desktop there isn't as critical of a noise and space issue. Having a near-line archive storage would be useful ( projects completed in last year (or so) that are archived for reference inside the Mac Pro. ). If Apple still have 3.5" drives in iMac 27" models in 2019/2020 then not a good reason to exclude them from Mac Pro. Driving the $/GB of capacity higher pushes away customers that are in the targeted market. $400 can get 10TB of HDD space which no SSD Apple is going to offer is even going to get in the same order of magnitude next to. Apple $800 jump from 1 to 2TB in SSD could get 20TB of HDD space. The capacity gap there is huge if mainly just want to store stuff nearby. Apple's sky high pricing is in part driving the need for HDDs in their solutions.
 
Apple seems to be conflicted with a "everyone skips into the future on APFS only SSDs" viewpoint on one hand and a "making affordable systems " viewpoint on the other. That's for desktops in general. Either Apple needs to fix their APFS ability to get along with HDD and hybrid systems or need to do a 180 turn on goosing SSD prices artificially high.
SSD prices have remained relatively static, if not a slight increase, over the past couple of years. If Apple is to offer lower cost, for Apple, SSD storage the overall cost of SSD technology needs to come down.
 
I think 10gbit ethernet is the modern equivalent of internal 3.5" drive bays. Data on spinning rust doesn't need to be internal any more. Put it in the server closet where you don't have to hear it, keep it cool, or power it with the workstation power supply. Even a giant multi-platter NAS isn't going saturate the network.

Generally this is true for most companies that have people working off a set of data that moved/shared between users. Even 5Gb/s ethernet (over decent wiring and some switch updates). Additionally TBv3 makes "sneaker net" even more viable in many cases ( e.g., a number of announcement at NAB of external storage enclosure than can take from 'field' to somewhere else. )

However, there are also buyers that are still in the "one man band" camp. Either as somewhat rogue operations inside of a larger company ( keeping some data storage and backup local) or just highly siloed work ( one person company or multiple folks with very low work data overlap). Apple probably needs the Mac Pro to cover different ground than what the iMac Pro covers ( and what the MP 2013 covered ).

3.5" spinning rust can be vastly different capacity if step out of the "what's the cheapest entry drive" zone though. 10TB is 10x bigger than a 1TB drive. That is what is being passed up.




Multiple M.2 slots (and enough PCI lanes to drive them) is the new target for workstations.

Not sure there will be multiple. Apple's boot SSD probably won't if the T2 is the groundwork future for workstations. One of those M.2 slots in mainstream ones is aimed at boot work. Apple could add one M.2 on the CPU PCI-e lanes. The iMac Pro has a x4 and x16 that are completely unused ( so just task that x4 to M.2. and the x16 to std PCI-e slot. ). Just getting Apple to admit that M.2 exist at all is a substantive change for them.

Going past one would probably mean sharing bandwidth. Pragmatically there is a switch in the PCH that could add another M.2 slot there. That bandwidth would have to be shared by the boot SSD though. That could be useful if it was a either/or context. That PCH hosted M.2 was a generally a Windows drive while the T2 boot SSD held just a macOS image ( or images via APFS ) . Getting Windows off the relatively limited capacity boot disk would actually be good for the Mac Pro. Two would be quite useful, but yet again it is getting Apple to even acknowledge M.2's existence is a hurdle.

That is typically how some mainstream boards have done it. One x4 link off CPU and then another off the PCH. ( probably based on premise that most folks are really going to add just one and will choose what the trade-off is going to make relatively the other components installed. )
[doublepost=1523821986][/doublepost]
SSD prices have remained relatively static, if not a slight increase, over the past couple of years. If Apple is to offer lower cost, for Apple, SSD storage the overall cost of SSD technology needs to come down.

I'm not talking about the panacea where folks several years ago where talking about how SSDs would drop to as low as $/GB pricing as HDDs by now. The issue is whether Apple is changing above market rates for SSDs. They are. I pointed to one example but there are many where Apple's upgrade prices and they prices for other SSDs on the market are no where near the same amounts.

Apple's 15" MBP. to jump from 256GB to 512GB is $200. To jump to 1TB is an additional $600 ( that is on top of paying for the 256GB SSD). Samsung PM981 (very recent ) and in limited OEM BTO only. Spot price:

".. We paid ~$200 for the 512GB PM981 and ~$400 for the 1TB model. .."
http://www.tomshardware.com/reviews/samsung-pm981-980-nvme-ssd,5323.html

Apple's 1TB SSD is greater than $600. Samsung SSD $400. Same base market for NAND chips. That's the issue. Apple almost prints money on their BTO options in the SSD space. As long as they heavily push that issue, HDDs will remain viable on the desktops. Pushing the SSD $/GB points artificially higher just make the value option that the HDDs have all that more viable. If Apple thinks that "HDDs are evil" they should be taking steps to get rid of them; not make them more viable.
 
Last edited:
  • Like
Reactions: -hh
3.5" spinning rust can be vastly different capacity if step out of the "what's the cheapest entry drive" zone though. 10TB is 10x bigger than a 1TB drive. That is what is being passed up.

Sure, but that's really parallel to my point which is simply that internal storage bays for hard drives is rapidly becoming less useful for this market segment. Hanging your slow storage off the network opens up many other benefits beyond just concurrent, shared access. You get filesystems designed for bulk storage, better redundancy options, and hot swap bays. Those are boring, useful attributes of bulk storage which (outside of the Xserve years) have been ignored by Apple and are likely to continue to be ignored.

10gbit ethernet really changes the dynamic, because there's no longer a performance penalty for moving your bulk storage outside of a workstation. The downsides persist but the upsides are dwindling.

Not sure there will be multiple. Apple's boot SSD probably won't if the T2 is the groundwork future for workstations.

I agree completely. I don't think the modular Mac Pro is likely to have anything as mainstream as an NVMe M.2 slot, much less multiples. For exactly the reasons you mention. It would be a great solution for Apple's customers but not for Apple. I was just sharing my perspective on the larger market. Despite Apple's strong "NIH" leanings, that's where their competition is headed and it's the landscape where the modular Mac Pro will land.
 
Last edited:
I know the argument has been made over and over again - I still believe Apple needs to stop being smart about the next MP's configuration and expandibility , and try their best to please everyone and serve every possible user case - within reason .
Kitchen sink with integrated bottle opener and corkscrew .

IMO, they need to return to a 'dumb' design just like the cMP , and on top of that work on third party compatibility .

Two things. First, what is 'reasonable' ( "within reason") is just like the listening issue. There are two sides that need to have a common understanding of the base words/concepts under discussion for there to be a shared understanding.

For lots of folks in these discussions "pro" means "what I and some close cohorts do" is professional and those other folks are not 'real' pros. If you throw "reasonable" into that kind of context then "reasonable" is "whatever feature I want". Chasing that kind of "everything and kitchen sink" is a problem. You are not really coming to a shared notion of "reasonable". It is basically a try to make everyone happy by offering everything.

That's not really a "dumb" design. That is really simply ignoring the issue. Being willfully ignorant and being "dumb" are two different things. There are very real issues with the Mac Pro and one of the primary ones is that the scale is going to significantly smaller than most folks think (when coming from "me super import, others don't matter").

As for the second issue, "third party compatibility". That is not magically decoupled from volume. The smaller market the 3rd parties see an opportunity, the less likely they are going to spend resources, money, and time trying to jump into the Mac Pro market. Chasing narrow niches can pay off in the general workstation because Dell and HP sell in the range of 800+ K workstations a quarter.

So if 800K/quarter workstation has a 2% +/- variable on that , a 3rd party vendor could see 16K oscillation uptick just tracking the larger market. Whereas if Mac Pro is less than 14 times less that that then the oscillation is only in the thousands range. If the 3rd party get 2% penetration into each one of those that's 320 for larger market and less than 23 for the much smaller one. In short, Apple listens to the 3rd party vendors too and if they say "that's too small for me to make money, I'll pass" that is very real too.

It is simply not true that number of slots is linearly going to get more users. 1-2 slots will probably cover the bulk. 3 would be only incrementally more. 4 only an even smaller incrementally more. An incremental niche of 2K folks may not be viable for a $30 card which only has $1-2/card to spend on driver development for that small a group.

What Apple needs to come up with is a reasonable number of slots in a new Mac Pro among PCI-e and drives. Some compromise, but not necessarily a complete flop of mindlessly copying the from of a HP Z6 or Z8 without any regard of how to fit into the rest of the Mac ecosystem.


That's what I'd consider 'listening' , and what well might be a major factor in winning back customers, and reassure existing ones .

There is also listening and being high priority. I'm sure Apple knows some folks won't like some of the design constraint choices.


Apple can only win with a highly flexible workstation design, but can not gain anything with any limitations in a future MP .

No limitation of any kind matches up with reasonable? Slapping a universal generalization on something usually isn't a concrete sign of being reasonable. that's typically done to overly simply the world as oppose to accurately describe it for what it is.


Apart from a smaller footprint and a few design awards , which obviously isn't working for everybody ...

The Mac Pro doesn't have to work for everybody. The whole Mac lineup doesn't have to work for everybody either. Apple has 7% of the classic PC market. How to optimize that is different from trying to manage 20-40% of the market.
 
  • Like
Reactions: vddobrev
Apple has made it abundantly clear the next MP will be as locked down and useless as it is possible to make; and will charge you exorbitantly in a bizarre inverse price - capability curve.
 
Actually, price is the only thing I’m worried about. I think the next MP will be just fine, but what if Apple feels that it’s base price should be higher than the iMac Pro?

I hope they keep the $3K base model, and I think anything more than $4K for the base model is gonna be insane.
 
  • Like
Reactions: askunk and Synchro3
Actually, price is the only thing I’m worried about. I think the next MP will be just fine, but what if Apple feels that it’s base price should be higher than the iMac Pro?

I hope they keep the $3K base model, and I think anything more than $4K for the base model is gonna be insane.
The 5k screen alone should shave off at least 1k if not more, so the rest of the components and the base spec that Apple decides to land on will determine the entry price. I would think it is in Apple's interest to price the base model competitively. Also if the design doesn't bring substantial benefits over other vendors' form factor then it will be a really hard sell, if it is not at least cheaper.
 
iMac Pro has almost workstation hardware and a 5k display, plus keyboard and mouse. I think the they 're going to price the 7,1 at comparable price and say: "Both models costs the same choose what you want..." hopefully.
 
Last edited:
I'm not talking about the panacea where folks several years ago where talking about how SSDs would drop to as low as $/GB pricing as HDDs by now. The issue is whether Apple is changing above market rates for SSDs. They are. I pointed to one example but there are many where Apple's upgrade prices and they prices for other SSDs on the market are no where near the same amounts.

Apple's 15" MBP. to jump from 256GB to 512GB is $200. To jump to 1TB is an additional $600 ( that is on top of paying for the 256GB SSD). Samsung PM981 (very recent ) and in limited OEM BTO only. Spot price:

".. We paid ~$200 for the 512GB PM981 and ~$400 for the 1TB model. .."
http://www.tomshardware.com/reviews/samsung-pm981-980-nvme-ssd,5323.html

Apple's 1TB SSD is greater than $600. Samsung SSD $400. Same base market for NAND chips. That's the issue. Apple almost prints money on their BTO options in the SSD space. As long as they heavily push that issue, HDDs will remain viable on the desktops. Pushing the SSD $/GB points artificially higher just make the value option that the HDDs have all that more viable. If Apple thinks that "HDDs are evil" they should be taking steps to get rid of them; not make them more viable.
Apple has always charged above market rates for upgrades and it is not something new. The reason I responded was to say if you expect to see a drop in $/GB for Apple's SSDs you'll have to wait for the overall market to drop.
 
iMac Pro has almost workstation hardware and a 5k display, plus keyboard and mouse. I think the they 're going to price the 7,1 at comparable price and say: "Both models costs the same choose what you want..." hopefully.
If the mMP starts at 5,000 USD then for me, the base model has to ship with components that exceed iMac Pro's spec, *and then* offering substantial room of expansion, wether it be 3rd party or Apple BTO. It is almost twice of what the 6,1 costs.
 
  • Like
Reactions: ssgbryan
Two things. First, what is 'reasonable' ( "within reason") is just like the listening issue. There are two sides that need to have a common understanding of the base words/concepts under discussion for there to be a shared understanding.

For lots of folks in these discussions "pro" means "what I and some close cohorts do" is professional and those other folks are not 'real' pros. If you throw "reasonable" into that kind of context then "reasonable" is "whatever feature I want". Chasing that kind of "everything and kitchen sink" is a problem. You are not really coming to a shared notion of "reasonable". It is basically a try to make everyone happy by offering everything.

Reasonable is but a word, I appreciate that, and common ground hard to find .
I'm using the term within the context of the discussions in this and the Waiting thread .
Also, I don't have the patience, nor good enough a command of the English language, to get every posting perfectly right . ;)

Big fan of your postings, btw . :)

As for the second issue, "third party compatibility". That is not magically decoupled from volume. The smaller market the 3rd parties see an opportunity, the less likely they are going to spend resources, money, and time trying to jump into the Mac Pro market. Chasing narrow niches can pay off in the general workstation because Dell and HP sell in the range of 800+ K workstations a quarter.
....

In short, Apple listens to the 3rd party vendors too and if they say "that's too small for me to make money, I'll pass" that is very real too.


I'm mainly talking about GPUs , RAM , internal storage and such ; products that exist and are perfectly usable out of the box .
In that area, Apple has been creating niches with proprietary solutions, not chasing them .
Arguably, that approach costs them more than using standardized parts, even figuring in expenses for GPU drivers / support, which they will likely share with the manufacturers anyways .




It is simply not true that number of slots is linearly going to get more users. 1-2 slots will probably cover the bulk. 3 would be only incrementally more. 4 only an even smaller incrementally more. An incremental niche of 2K folks may not be viable for a $30 card which only has $1-2/card to spend on driver development for that small a group.

What Apple needs to come up with is a reasonable number of slots in a new Mac Pro among PCI-e and drives. Some compromise, but not necessarily a complete flop of mindlessly copying the from of a HP Z6 or Z8 without any regard of how to fit into the rest of the Mac ecosystem.

I strongly disagree, with the bolded part in particular .

One slot more than whatever imaginary percentage of users will ever need - doesn't matter, doesn't cost much, doesn't need much space .
Most importantly - no complaints .
I put that reasonable number of slots at 4 .

Too few slots ( assuming there will be PCIe slots to begin with ) - and the next MP will be percieved as restricive and uncompetitive on even the most basic level .
Keep in mind that the MP will most likely come in one version only, unlike HPs and such which can offer different models with varying degrees of expandibility .


There is also listening and being high priority. I'm sure Apple knows some folks won't like some of the design constraint choices.

I'm fairly certain Apple only have the faintest idea on just how many folks are fed up how much with their recent Mac design choices .


No limitation of any kind matches up with reasonable? Slapping a universal generalization on something usually isn't a concrete sign of being reasonable. that's typically done to overly simply the world as oppose to accurately describe it for what it is.

The Mac Pro doesn't have to work for everybody. The whole Mac lineup doesn't have to work for everybody either. Apple has 7% of the classic PC market. How to optimize that is different from trying to manage 20-40% of the market.

No limitation, in the context of this discussion .
Again, GPUs, storage, Ram, PCIe slots, ports, things like that .

I do believe Macs have to work for as many people as possible .
Especially because of their low market share - they can't afford to become even more niche by further limiting the hardware side of things .

This isn't the 90s anymore, when the PC market was developing, being different was a virtue , and there was still a chance of having multiple competing OSs .
[doublepost=1523880391][/doublepost]
I think 10gbit ethernet is the modern equivalent of internal 3.5" drive bays. Data on spinning rust doesn't need to be internal any more. Put it in the server closet where you don't have to hear it, keep it cool, or power it with the workstation power supply. Even a giant multi-platter NAS isn't going saturate the network.

Multiple M.2 slots (and enough PCI lanes to drive them) is the new target for workstations.

I completely forgot about the server closet ; got to call the IT guy to tell me where to put the new enclosure . ;)

Kidding aside, internal storage is still a thing ; most of my files are external, but having a bunch of current stuff on high capacity HDDs inside my cMP is very convenient .
Especially since I'm my own IT guy and the server closet is a ratty table in the next room .
 
iMac Pro has almost workstation hardware and a 5k display, plus keyboard and mouse. I think the they 're going to price the 7,1 at comparable price and say: "Both models costs the same choose what you want..." hopefully.

Yup, the iMac Pro is a $5k workstation, with a free display. The Mac Pro will be a $5k workstation with free expansion / upgrade capacity. Pricing will be cpu / gpu / storage / memory equivalent, just one of them will be able to go higher capacity-wise (hopefully).
 
  • Like
Reactions: filmak
If the mMP starts at 5,000 USD then for me, the base model has to ship with components that exceed iMac Pro's spec, *and then* offering substantial room of expansion, wether it be 3rd party or Apple BTO. It is almost twice of what the 6,1 costs.
Yes, I agree.
It is really too much, I would like it to start at ~2,500, but it is up to Apple to decide what to charge for the extra space needed for a couple PCIe slots.
 
  • Like
Reactions: ETN3
While we're discussing pro pricing - iMac Pro base model is still discounted at major retailers. This bodes well for a reasonable point of entry.

Why is it still discounted.. within months of launch ?

Anyway.. if Apple can release the base nMP with 10 core Xeon, 64 GB Ram, vega 64 ( or equivalent), 1 TB and at least two open x16 PCI-e slots under 5k USD ... it just might come out of the gates swinging. Might.
 
Anyone happen to know the maximum theoretical length for a pci riser cable? Looking at a pc I’m thinking of building, it occurred to me that eGPU doesn’t necessarily require thunderbolt, eg magma’s chassis back in the day...

A Mac Pro with no internal card slots (or the space they use), but, say 3 or 4 dedicated “pci direct” ports that are just a full-fat pci slot reshaped into a plug connector, which support a single device each, could be a workable compromise. Put gpu (or decklink 8k) into box, plug whatever into card, plug box into “slotport” or “magicbus” or “cyclone” or whatever goofy name marcom comes up with... “slotless” Mac Pro with expansion capabilities that aren’t maxed at 1/4 the lanes of traditional workstations.

Anyone? Anyone? Bueller?
[doublepost=1523974337][/doublepost]

Guarantee you, that’s how they’ll position it. Put it this way, if the pro display they’re developing is the same as the one in the iMac Pro*, the price of a Mac Pro plus the pro display, will be the same as a similar config iMac Pro with a (second) pro display.

*my money is still on the pro display being 32”8k, and damn the number of cables it requires.
 
Last edited:
Anyone happen to know the maximum theoretical length for a pci riser cable? Looking at a pc I’m thinking of building, it occurred to me that eGPU doesn’t necessarily require thunderbolt, eg magma’s chassis back in the day...

Look into
https://www.cubix.com/xpander-desktop-elite-gen-3-1500w/
And
https://www.onestopsystems.com/categories/desktop-expansion
And
http://www.netstor.com.tw/product_info.aspx?PID=PID_170324287237937

eGPU is an intel standard to push GPUs over TB3.


[doublepost=1523984279][/doublepost]
*my money is still on the pro display being 32”8k, and damn the number of cables it requires.

Apple might release both 5k 27” displays and maybe a 32/40 inch 8k variety, provided the latter is allowed to run at 60hz, 10 bit, uncompressed over HDMI 2.1, or the next display port standard.
 
Last edited:
Yup, the iMac Pro is a $5k workstation, with a free display. The Mac Pro will be a $5k workstation with free expansion / upgrade capacity. Pricing will be cpu / gpu / storage / memory equivalent, just one of them will be able to go higher capacity-wise (hopefully).

For this kind of service / support, it's scary to treat the iMac Pro as a real workstation.

 
That is exactly why I will not purchase an "all-in-one" as a pro workstation. It's bad enough that their laptops are barely repairable. The iMac Pro looks cool on a stage or at a conference for a demo, but that's not where I'd be using it 99% of the time.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.