MP 7,1 Waiting for Mac Pro 7,1 A1991 (no more)

[deconstruct60]

sata has had hot swap for some time. We do not need file system level raid. We need block level raid Does not mac os not have mdraid?? (does not need to be boot disk)
Also don't force people into bios fake raid

APFS is not just a file system it subsumed both Core Storage and HFS+ . Fusion Drive is block level solution (the new APFS provisioned on will likely be one too). There is no "BIOS" that Apple like to incorporate into their solution at all. If trying to talk about software boot support nobody is being forced into that at all. If folks want to buy stuff and add it to their new Mac Pro later that would still be an option. Nor does it "have to be" a bootable solution. However, the Apple configurations that they sell just probably won't incorporate that. Nor will Apple likely sell any "real RAID card" in their store. It isn't a focus nor absolutely critical to the market as long as they provide slot for folks to plug in their "real RAID card", but those don't necessarily have to point at internal drives in Apple sleds.

SATA having hot swap is immaterial. Something up the stack from SATA has to make make that disk that disappearance doe not appear to the file system. The point is that Apple isn't going to be selling that. The Mac Pro 2006 with the custom RAID card is not where Apple is likely to go back to. The Mac Pro from 2007-2008-2009-2010 evolved to increasingly less aligned with internal redirection of the drive sleds to a "real RAID card". Again, it is highly unlikely that Apple is going to deviate from that evolutionary path.

also why not have e-sata??? it's free with the chipset and does not tie up the TB / USB bus. (other then the DMI bus)

Other than sunk costs in eSATA infrastructure, it is more performant than Thunderbolt v3 how? Like the MP 2013, the next Mac Pro likely will not put the TB bus behind the DMI bus. ( the iMac Pro doesn't either).

Given that the current x4 PCIe-v3 SSD are pretty close to saturating the DMI bus, that's probably a contributing reason why Apple isn't particularly keen on striped SATA solutions. If you put two x4 PCI-e SSDs on the PCH and actually use them at the same time then then can most definitely saturate the DMI bus. (e.g., Apple puts an empt M.2 slot on the PCH for an optional drive ).

If look at the iMac Pro block layout, they moved the 10GbE chipset off the PCH ( and DMI) path altogether; it hangs off the CPU lane block.

If Apple puts 4+ USB Type-A 3.1 gen 1 (or 2 if a new model takes even more than a year ) sockets on the next Mac Pro and provisions off the "already paid for" PCH then that will probably add to DMI bus traffic. And external SSD and T2 SSD in concurrent use DMI basically done. Apple is far more likely to use the 'free' high speed USB functionality of the PCH, than the SATA. That's another 5-10 Gbps.


if look in the general Windows PC workstation market the standard configurations are typically single boot HDDs which are not threat to the DMI bus saturation or perhaps some SATA SSD which are "throttled" back to SATA limits. (again no threat to DMI saturation). For the Mac Pro workstation market it is extremely likely going to be the exact opposite approach. The lowest configuration boot storage device is going to a blazingly fast SSD attached to the PCH which is relatively close to saturating the DMI bus.

At $3+ K priced systems Apple isn't going to be trying to sell these to folks who want to go affordably fast; it is performance for money class.

eSATA is more so a sunk cost pit that Apple could consider bumping the empty PCI-e standard slot count closer to 2-3 than perhaps just one. ( e.g., put a PCI-e switch on the link the 10GbE controller(s) is on and weave in a x4 PCI-e electrical , x8-16 physical slot)

 

[-hh]

https://www.newegg.com/Product/Prod...r=1&cm_re=hades_canyon-_-56-102-201-_-Product

An interesting alternative to a 'pro' (ish) mini.

Looks like a basic deck-out at 32GB RAM & 0.5TB M.2 with Windows OS is roughly $1.5K. The alternative from Apple is a 21.5" iMac (4K) where the rationalization is that the 21.5" display costs $1000.

Considering that I already have a 27" Apple LCD Display, that extra $1K doesn't strike me as a particularly wholesome value....

 

[danwells]

There is no "lower end of the Mac Pro market", not one that Apple is interested in, anyway. The Mac Pro is a "machine of last resort" Apple sometimes sells to people with specialized needs no mainstream Mac will meet. They have zero interest in offering an expandable (or even headless) Mac with mainstream performance. The fact that there is an odd low-end iMac using a 15 W mobile CPU that performs in Mini territory doesn't change the fact that iMac territory (all standard desktop processors) has contained nothing but iMacs for twenty years. A few odd iMacs reaching below their standard territory, and, with the iMac Pro, one reaching substantially above, doesn't change the fact that they occupy their territory exclusively.

The iMac's exclusive territory (assuming the iMac Pro is not a protected iMac) puts the bare minimum starting price of the new Mac Pro at $3699 plus the 10-20% premium the Mac Pro has always had. Yes, exotic BTO iMacs have sometimes reached into Mac Pro territory, but I am almost sure that a 27" with 32 GB of RAM and a 1 TB SSD is not an exotic BTO, if for no other reason than that the Mac Pro will start with at least those specs.

I would expect the Mac Pro to start substantially higher than that, based on Apple's comments on throughput. I don't think it's an iMac Pro in a different case - it's something more than that, and the most obvious thing it could be is Xeon-SP. If it is a Xeon-SP machine, the starting price is likely to be around $6499 (or more, depending on how high the RAM, SSD and GPU go on the base model).

Bloomberg's comments on a new Mini being of more pro interest could mean as little as "there will be a 28 W mobile quad-core as a BTO option", or it could mean something more. The most likely "something more" is that it's still a headless laptop, but a more powerful one (could be anything up to and including the mobile i9/Radeon 560 combination from the 15" MBP as a top option). This would require a new case with more cooling, and any hex-core option would require a simultaneous update to the iMacs to keep them comfortably ahead - current desktop chips would do it, and the new ones coming on October 1 would produce a larger lead.

I know that others disagree, and I don't think it's especially likely, but I could see three desktops (not counting iMacs) in the upcoming lineup. One is, of course the Mini, and another the modular Mac Pro (if this happens, the modular Mac Pro is not even a $6499 machine, but more of a $10,000+ machine, and the Mini doesn't get a 6-core option). The third would be a not terribly expandable screenless version of the iMac Pro. It would be more of a successor to the trashcan than anything else. The constituency is photographers and video users (below the extremely high end occupied by the modular Mac Pro) who use Adobe RGB monitors and the like. It would certainly have expandable RAM, maybe a slot or two for a PCIe SSD, but either no standard PCIe whatsoever or at most one slot with limited length and power for a RAID controller or a music interface.

This doesn't alter Apple's stance that the iMac is the mainstream desktop - it still starts at $3699 or above - and it doesn't bring a slotbox within reach of people who mainly want a GeForce in their Mac (which Apple really wants to avoid). It provides a second Xeon-W/Vega alternative to the iMac Pro. Is this a super Mini, or is it a Mac Pro?

If they did this (going way back in Mac history, the relationship between this machine and the modular Mac Pro is the relationship between the IIci and the IIfx), the modular Mac Pro would be extremely high-end. It would certainly be dual CPU capable (maybe even dual CPU only), and Apple might find some odd way of making it quad CPU capable (the Xeon-SP chips they might use are all quad-capable). They couldn't just throw four sockets in the box - serving an exotic configuration that would account for 10% of sales at most would mean the machine would no longer plug into a standard 120V outlet. Xeon-SPs are up to 250 watts each, add in the power draw of the rest of the machine (including multiple high-power GPUs and accelerators), and you'd need something like a 2500 watt power supply. A 20 amp 120V outlet only supports a (roughly) 1600 watt power supply at most.

The only way I could see Apple pulling off a quad CPU option is to extend the UPI links between processors to a connector that allowed a second dual-processor module outside the main case (probably attached to the side or back, not cabled - those links have to be kept short). The additional processor module would have a 240V, 20 or 30 amp power connector that then split the power between the main power supply and its own power supply.

Again, I'm not saying Apple will do any of this, just that it's possible, and perhaps more possible than the $2999 slotbox that generates so much forum interest. Apple knows about the forum interest, has decided it's a small minority of users (and would be a pain in the ass to support), and has a deal with AMD (not to mention that their own software, especially Final Cut, runs best on a Radeon). They also know that games and gaming-focused graphics cards are not the most stable things around. Certainly not everyone who wants a GeForce wants it for gaming, but it's part of the appeal, and it's a part Apple has tried hard to steer clear of.

 

[AidenShaw]

Instead of a nice, clean interior layout (something the cMP case was praised for) one now has a rats nest of cables (both data and power).

And few of the myriad cables have latching mechanisms - so any disturbance risks unplugging important external devices (and the risk of loss of data).
[doublepost=1535144293][/doublepost]

https://www.newegg.com/Product/Prod...r=1&cm_re=hades_canyon-_-56-102-201-_-Product

Thanks. Interesting that Intel is upping the ante on the NUCs.

But if you want a really interesting mini - of course it's in the Z-series line. http://www8.hp.com/us/en/workstatio...n/psg/ws_desktops/products/z2-mini-learn-more

Xeon, ECC RAM, and Quadro Pascal GPU....

And HP has clever engineering and marketing (maybe they follow MacRumours).

​

Note that, since this sub-tangent is speculating about an unannounced MiniMac update, this HP link is for the *announced* but not yet shipping Z2 Mini G4. The current G3 Mini is similar, but has an earlier generation Xeon and Maxwell graphics.

 

[Biped]

But if you want a really interesting mini - of course it's in the Z-series line. http://www8.hp.com/us/en/workstatio...n/psg/ws_desktops/products/z2-mini-learn-more

Xeon, ECC RAM, and Quadro Pascal GPU....

And HP has clever engineering and marketing (maybe they follow MacRumours).

You know, if we all open safari ( for once ) and load up these HP Z-series workstation links in it. Maybe we can cause some panic at Apple WHQ.

 

[deconstruct60]

You know, if we all open safari ( for once ) and load up these HP Z-series workstation links in it. Maybe we can cause some panic at Apple WHQ.

Browser histories is an extraordinarily poor way to drive market research on a specific product. People are largely 'reading' in browsers not 'doing'.

Second, Apple doesn't use market focus groups as the primary driver of their marketing. Again people can mainly tell you what they know ( or think they know), but not so much what they could do with something they have never used.

Third, Apple probably has someone ( or group) doing competitive analysis. If some "insanely great' feature pops up they may 'steal' it ( as in great "artists steal' not so much as slavishly copy), but Apple has about zero interest in some sort of feature and product checkbox 'war' of matching Dell/HP/Lenovo across all features and all products. The strategic policy is to do a limited number of chosen products 'better'. A "monkey see , monkey do" is about 180 degrees opposite of their approach. Slavishly tracking browser histories is pure "monkey see, monkey do" tactics.


That's said, it is kind of humorous to have seen all the rants about how "Nobody asked for a smaller workstation" hurled at the Mac Pro 2013 and then point to the HP mini as a 4th generation product. ( If nobody is buying these why is HP on the 4th iteration.... wouldn't they have stopped by now if there were no buyers and the previous three iterations had all been big money losers? ) The market is no where near almost completely homogeneous in slavishly worshipping at the commodity GPU card form factor first alter as most of the protesters try to make it out to be. it is a different market space than the 2006 era Mac pro was in but it isn't empty if using the current technology available.

 

[Biped]

It was meant as a quip.

noun

1. a witty remark.
   synonyms: joke, witty remark, witticism, jest, pun, bon mot, sally, pleasantry;
   

That aside, I agree with your remarks, there is too much 'no true scottsman'-ship generally in these forums ( grant me license to stretch the meaning ). People assume their usage cases are often the preeminent case. Or fail to realize that they might not be able to imagine what happens in environments outside their own. You get a sort of dissonance between their own reality and one often different supported by tangible evidence, such as market indicators, etc ...

 

[AidenShaw]

You know, if we all open safari ( for once ) and load up these HP Z-series workstation links in it. Maybe we can cause some panic at Apple WHQ.

... and then use Apple Pay to buy the Z-series

 

[barmann]

That said the next Mac Pro should have more than one SSD. SSDs get performance, but are more slowly coming down the tracks in affordable capacity. It is getting better, but not as fast as the hype from several years ago about SSDs catching HDDs in a couple of years.

Amen .

Another thing worth considering is form factors and computer specific drives .

M.2 drives - the ones faster than 2.5" - might get closer in price to 2.5" drives soon, and more and cheaper external and internal solutions are likely to pop up soonish as well, but we're not quite there yet .

Let's keep in mind that there is a point of diminishing returns - faster storage is always better, but in many use cases the extra cost and a less flexible form factor can not only mean they are unnecessary, but even a liability .

It might be what you consider another irrelevant subset of customers, but there are a bunch of people in the professional world who throw harddrives around like playing cards, and move them from one computer or enclosure to the next, store them in vaults, mail them around the world etc ..
For that purpose, 2.5" SATA SSDs - and even spinners - are the new thumbdrive .

Also, drives dedicated to a single computer, like Apple's T2 approach, are just a waste of valuable space in that respect .
Or in any other respect .

 

[deconstruct60]

Amen .

One SSD not being a sole solution doesn't rule out more than one SSD being substituted. While the $/GB is behind the aggressive "analyst' forecasts they are getting close to be HDD prices were several years ago (back when folks were buying multiple drives to get things done .... *cought* like the 200x series Mac Pros. ).

Another thing worth considering is form factors and computer specific drives .

M.2 drives - the ones faster than 2.5" - might get closer in price to 2.5" drives soon, and more and cheaper external and internal solutions are likely to pop up soonish as well, but we're not quite there yet .

Probably not. Mainstream SATA probably isn't going to move foreward from the 6Gb/s limit. So the SSD ontrollers for SATA drives can be cheaper because they don't have to go faster than 6Gb/s. Similarly, cheaper, slower write QLC solutions are a better fit to having the interface capped at a lower speed.

"slow" m.2 are possible. But the issue with M.2 is the limited space on the blade. The 2.5" has the "problem" that it is grossly wasteful of space ( the actual SSD inside is often a significantly smaller subset of the volume available. ) This is where the ultra-high cost, "mega" capacity SAS drives are going, fill up a much larger fraction of the internal volume with NAND chips. 2.5" drives will be able to "go wide" (use more NAND chips) of mature densities (older) while the M.2 designs have to maximize density to get to max capacity.

Where there is flexibility in the enterprise space see similar issues ( plus supercapacity space to shut down cleaning. ).

Intel's 'Ruler" drives
https://www.anandtech.com/show/11702/intel-introduces-new-ruler-ssd-for-servers

More rulers and bigger rectangles.
https://www.anandtech.com/show/13218/ssd-form-factors-proliferate-at-flash-memory-summit-2018

Let's keep in mind that there is a point of diminishing returns - faster storage is always better, but in many use cases the extra cost and a less flexible form factor can not only mean they are unnecessary, but even a liability.

There are also liabilities in keeping archival, low access data spun up on near-line, "fast" storage too.

It might be what you consider another irrelevant subset of customers, but there are a bunch of people in the professional world who throw harddrives around like playing cards, and move them from one computer or enclosure to the next, store them in vaults, mail them around the world etc ..
For that purpose, 2.5" SATA SSDs - and even spinners - are the new thumbdrive .

Raw, bare 2.5 SSDs ? Not really. SSDs in external drives, enclosures , and portable "plug-and-play" containers into enclosures. yes. But that is actually part of the point of Apple's move to single boot drive. IF lots are data is being handed around via sneaker-net then the storage is external.

The uptake of SAN, NAS, and DAS is growing out there. Affordable 10GbE will only make the uptake on SAN get bigger as more workloads "just work".

The subset that Apple isn't placing at high priority is the "packrat" crowd. The folks in the camp of "Every project I've worked on for the last 5-15 years has to fit inside my one machine" crowd. It is the single lone wolf who works outside of local work groups and sharing is a low priority activity.

Also, drives dedicated to a single computer, like Apple's T2 approach, are just a waste of valuable space in that respect .
Or in any other respect .

The T2 doesn't use any significantly more space with the same NAND chips than the others solutions do. It is aimed at least as much toward security as much as speed. And it is basic fallout of Apple defacto being in the SSD making business. Fighting over the boot drive is a waste of time. The core issue for Apple is that they need to acknowledge that there are other 3rd party SSDs out there. ( e.g., hiding TRIM by default from non Apple drives is completely misguided at this point. At one point they may have had a lame excuse but now it is beyond lame to have that attitude. )

 

[Mago]

Just saw an TB3 external das case from OWC for 4 m.2 nvme ssd, compact, upto 8TB and 2800MB/s transfer.

 

[AidenShaw]

Just saw an TB3 external das case from OWC for 4 m.2 nvme ssd, compact, upto 8TB and 2800MB/s transfer.

OWC bay ($350) + 4*2TB 970 EVO ($700 each) = $3150 for 8 TB
Seagate 8TB Ironwolf SATA (to connect to internal SATA bay) = $244 for 8 TB

If you need 8TB with 2.8 TB/sec transfer - that's a nice match.

If you need 8TB for archival storage, the $244 Seagate is good (as long as your system isn't crippled by having no SATA slots).

With Z, Precision or other workstations - you could pick "A" or "B".

 

[Lammers]

Also, drives dedicated to a single computer, like Apple's T2 approach, are just a waste of valuable space in that respect .
Or in any other respect .

They do, however, contribute meaningfully to Apple’s profit margin, which is an important consideration in their design decisions.

 

[-hh]

...
It might be what you consider another irrelevant subset of customers, but there are a bunch of people in the professional world who throw harddrives around like playing cards, and move them from one computer or enclosure to the next, store them in vaults, mail them around the world etc ..
For that purpose, 2.5" SATA SSDs - and even spinners - are the new thumbdrive .

Raw, bare 2.5 SSDs ? Not really. SSDs in external drives, enclosures , and portable "plug-and-play" containers into enclosures. yes. But that is actually part of the point of Apple's move to single boot drive. IF lots are data is being handed around via sneaker-net then the storage is external.

Sometimes it may be in an enclosure, other times not...but that's not really the point:

The point is that this 'Sneaker Net' approach is still very much used in some customer markets in lieu of modern Cellular/WiFi bandwith connections. Even though there's bigger & faster pipes today, data consumption demands have grown too. For example, a decade ago, a dSLR was 6-8MP and usually shot JPEG...but today, they're 20+MP and shooting RAW, so the demand has grown by an order of magnitude (from ~3MB/image to ~32MB/image).

The subset that Apple isn't placing at high priority is the "packrat" crowd. The folks in the camp of "Every project I've worked on for the last 5-15 years has to fit inside my one machine" crowd. It is the single lone wolf who works outside of local work groups and sharing is a low priority activity.

"Packratting" is a different dimension to this.

The uptake of SAN, NAS, and DAS is growing out there. Affordable 10GbE will only make the uptake on SAN get bigger as more workloads "just work".

True, but it still depends on who-what-where.

At "home", with ISPs becoming more draconian on limiting bandwidth consumption, the prospects of shuffling around a couple of TB through their network for making a backup (or restoring) invokes interest in bringing the high bandwidth consumption task to be more closely in-house than out-of-house...not every small business is going to have the budget for a local 10GB NAS/DAS, particularly if the majority of the data pushing is from on/few specific individuals.

Similarly, price-performance points inform workflows & change paradigms. Much like USB thumb drives, the price of a 1TB 2.5" drive is only $50 today, so while it won't be literally given away at trade shows (I've gotten 4GB Thumb drives this way), for collaborating colleagues to facilitate bulk transfers, its viable...indeed, being done, because its {faster/easier/better} than burning a pile of Optical CDs.

And even for the Prosumer - consider the historical price gradient for camera memory cards - - its price per GB is an order of magnitude less expensive than ~15 years ago. As such, instead of hauling along a laptop (or even the short-lived "digital wallet" technology), one can functionally obviate the need to spend touch labor while still on the road "managing data" by just buying a ton of memory cards: the cost today for 300GB is only $200.

But why take this sneakernet-esque approach? Because having ubiquity of the internet doesn't automatically mean bottomless (& fast) bandwidth. Personally, I've been places on holiday and discovered that when the tourists return to the hotel each afternoon, all of their iPhones all simultaneously try to synch the day's pics, promptly bringing the host network to its knees. And thanks to streaming movies, hotel policies are increasingly capping customer bandwidth - - even on overnight overnight transfer strategies. And sure, you can walk down the street to the local Cyber Café ... but that's not a particularly productive use of your time to sit there and wait for 50GB to upload. What's "best" for productivity changes with your workflow context.

 

[fuchsdh]

But why take this sneakernet-esque approach? Because having ubiquity of the internet doesn't automatically mean bottomless (& fast) bandwidth. Personally, I've been places on holiday and discovered that when the tourists return to the hotel each afternoon, all of their iPhones all simultaneously try to synch the day's pics, promptly bringing the host network to its knees. And thanks to streaming movies, hotel policies are increasingly capping customer bandwidth - - even on overnight overnight transfer strategies. And sure, you can walk down the street to the local Cyber Café ... but that's not a particularly productive use of your time to sit there and wait for 50GB to upload. What's "best" for productivity changes with your workflow context.

All these latter points are arguments for external hard drives, not really anything about a professional workstation.

 

[ssgbryan]

All these latter points are arguments for external hard drives, not really anything about a professional workstation.

I love how some of you change the goal posts as soon as a real world example is provided.

A "professional workstation" is used by whoever needs that power: amateur, prosumer, or professional.

I am not using a cMP because I am a "professional" - I use the cMP because it is the only computer apple makes that won't melt when I start to render out my artwork.

In the 3d art world, even the most inexpensive (or free) software will bring any iMac to it's knees, and yes, that also includes the iMac "Pro".

 

[barmann]

"slow" m.2 are possible. But the issue with M.2 is the limited space on the blade. The 2.5" has the "problem" that it is grossly wasteful of space ( the actual SSD inside is often a significantly smaller subset of the volume available. ) This is where the ultra-high cost, "mega" capacity SAS drives are going, fill up a much larger fraction of the internal volume with NAND chips. 2.5" drives will be able to "go wide" (use more NAND chips) of mature densities (older) while the M.2 designs have to maximize density to get to max capacity.

Where there is flexibility in the enterprise space see similar issues ( plus supercapacity space to shut down cleaning. ).

Intel's 'Ruler" drives
https://www.anandtech.com/show/11702/intel-introduces-new-ruler-ssd-for-servers

More rulers and bigger rectangles.
https://www.anandtech.com/show/13218/ssd-form-factors-proliferate-at-flash-memory-summit-2018

https://www.anandtech.com/show/13218/ssd-form-factors-proliferate-at-flash-memory-summit-2018

That looks interesting, the ruler concept .
But the issue might rather be pricing, as you mentioned ; 2.5" SSDs are based on the old HDD standard obviously . And like you said, they have space for more chips, yet the issue remains to be affordability .

There are also liabilities in keeping archival, low access data spun up on near-line, "fast" storage too.

True . I'm not suggesting rarely used archives need to fit inside a workstation .

Raw, bare 2.5 SSDs ? Not really. SSDs in external drives, enclosures , and portable "plug-and-play" containers into enclosures. yes. But that is actually part of the point of Apple's move to single boot drive. IF lots are data is being handed around via sneaker-net then the storage is external.

The uptake of SAN, NAS, and DAS is growing out there. Affordable 10GbE will only make the uptake on SAN get bigger as more workloads "just work".

The subset that Apple isn't placing at high priority is the "packrat" crowd. The folks in the camp of "Every project I've worked on for the last 5-15 years has to fit inside my one machine" crowd. It is the single lone wolf who works outside of local work groups and sharing is a low priority activity.

Handling bare SSDs is not common in general, but not uncommon on Film and Photo shoots or other productions .
Mostly we are talking basic USB 3.0 and some occassional TB external drives though , that's true .

However, for the kind of data transfer I was talking about, the majority of drives used are 2.5" and some 3.5" ones . That form factor isn't going anywhere anytime soon, and not only in certain industries .
And that form factor is what I believe should have its space in a workstation as well, along with more current solutions .
Similar say, to a cMP ...

As for networked and other local storage, that works for many scenarios .
A lot of high volume data transfer outside a sophisticated local ( as in the building local + IT guy with ponytail ) infrastructure is very much done by sneaker-net ( is that a term ? ) and FedEx .

Apart from that - and this has been discussed to death - fast, quiet, reliable external storage enclosures cost a bunch of money .
I'm just talking about data I need to have on hand constantly, not the stuff I keep on external drives that only run when necessary .

And I want and need it to be on seperate drives, for redundancy, speed (scratch etc.) , convenient replacement and upgrades . Its under 2TB in total . Plus a 4TB HDD for my movies, just because I can .

Bigger companies with different storage setups, people who have basic requirements, they might have different needs .
But noone wants to pay Apple for the 2TB upgrade of a single internal SSD , when they used to get a BTO build with an extra drive for a lot less , or the IT guy/ nerd friend did it for even less .

The T2 doesn't use any significantly more space with the same NAND chips than the others solutions do. It is aimed at least as much toward security as much as speed. And it is basic fallout of Apple defacto being in the SSD making business. Fighting over the boot drive is a waste of time. The core issue for Apple is that they need to acknowledge that there are other 3rd party SSDs out there. ( e.g., hiding TRIM by default from non Apple drives is completely misguided at this point. At one point they may have had a lame excuse but now it is beyond lame to have that attitude. )

I mostly agree; my point is rather that a dedicated boot drive is a waste of space because that space could be used for a user replaceable solution .

 

[fuchsdh]

I love how some of you change the goal posts as soon as a real world example is provided.

A "professional workstation" is used by whoever needs that power: amateur, prosumer, or professional.

I am not using a cMP because I am a "professional" - I use the cMP because it is the only computer apple makes that won't melt when I start to render out my artwork.

In the 3d art world, even the most inexpensive (or free) software will bring any iMac to it's knees, and yes, that also includes the iMac "Pro".

I'm not moving any goalposts. I haven't set any goalposts to begin with, so maybe reread instead of jumping up to post without thinking. The rest of your post is just hyperbole, so it's not even worth responding to.

 

[danwells]

8TB 7200 RPM enterprise drive (probably that same Ironwolf from the example above) in USB-C external case: $299
https://www.bhphotovideo.com/c/prod...ologies_bbpr8000_8tb_blackbox_pro_rugged.html

That's why Apple won't put a big ol' hole for it in the Mac Pro. If you want a bunch of 'em, you can toss 'em in here...
https://www.bhphotovideo.com/c/prod...ctb3ivkit0gb_thunderbay_4_four_bay_drive.html

There are other, similar products... There is no configuration of hard drives that benefits from being internal - if you have enough of them to saturate a cheap USB port, you would need four or more bays in the case - and that adds cost and bulk (more than buying a Thunderbolt enclosure that keeps up with 4-8 drives). If you have too many for Thunderbolt, they'd never fit in any conceivable case (except, perhaps for a rackmount storage server).

 

[barmann]

That's why Apple won't put a big ol' hole for it in the Mac Pro. If you want a bunch of 'em, you can toss 'em in here...
https://www.bhphotovideo.com/c/prod...ctb3ivkit0gb_thunderbay_4_four_bay_drive.html

There are other, similar products... There is no configuration of hard drives that benefits from being internal - if you have enough of them to saturate a cheap USB port, you would need four or more bays in the case - and that adds cost and bulk (more than buying a Thunderbolt enclosure that keeps up with 4-8 drives). If you have too many for Thunderbolt, they'd never fit in any conceivable case (except, perhaps for a rackmount storage server).

Additional internal drives are benifitial in any workstation ; I'm baffled by your claim .

Alright then - how much cheaper would a cMP have been with just one drive bay , and without optical drive bays ( I'm all for not having the latter ) ?

Will Apple deduct the price of an external harddrive enclosure off the price of the next MP, if it has less than say 4 drive bays of any kind ? Surely any savings will be passed on to the customer .

How do I add a fast drive(s) for scratch, program libraries, heavy I/O work, temporary project files etc . ?
TB is often too slow ; should I just get the Apple upgrade from their single base, proprietary 512GB SSD to 2TB for 1299 + tax , and pray it never breaks ?

Do I have to have an extra box around me for just a couple of essential drives , which might or might not be fast and quiet enough, and the intern or the neighbour's cat won't disconnect twice a day ?

 

[ssgbryan]

That's why Apple won't put a big ol' hole for it in the Mac Pro. If you want a bunch of 'em, you can toss 'em in here...
https://www.bhphotovideo.com/c/prod...ctb3ivkit0gb_thunderbay_4_four_bay_drive.html

There are other, similar products... There is no configuration of hard drives that benefits from being internal - if you have enough of them to saturate a cheap USB port, you would need four or more bays in the case - and that adds cost and bulk (more than buying a Thunderbolt enclosure that keeps up with 4-8 drives). If you have too many for Thunderbolt, they'd never fit in any conceivable case (except, perhaps for a rackmount storage server).

Do you use that product? Do you use any similar product? I have it's ESata predecessor - I like it, but that doesn't mean that I want a couple of them across my desk, each needing it's own electrical outlet, and adding 3 additional points of failure for each enclosure.

You do understand how much more you are adding to the TCO, do you not? Dumping the internal drives made the Trashcan even less competitive with it's counterparts from HP and Dell. In my case, it would have cost me an additional $1,200.

I am always baffled by folks that don't have large amounts of data telling those of us that do what our requirements are.

 

[Chancha]

I feel that the recent pages of drive centric discussion was caught up between what we (this thread's demographic) wants, vs what we expect Apple to ship.

(traditional) workstation: of course internal drives is needed if not necessary, so that any attempt to modularize storage externally a la tcMP is just a waste of resource. Unused drive slots and ports on mobo are much less of a problem than dangling AC bricks and TB cables.

(what Apple conceives) modern "pro workflow": independently modular components that work and power on its own, inter-connected via sub-optimal I/O that otherwise enables flexibility (eg. eGPU that can stay home while you take MBP out, or TB RAID enclosure that is not bound to this and only this tower)

The question is where Apple is positioning itself within this spectrum.

 

[barmann]

I feel that the recent pages of drive centric discussion was caught up between what we (this thread's demographic) wants, vs what we expect Apple to ship.

(traditional) workstation: of course internal drives is needed if not necessary, so that any attempt to modularize storage externally a la tcMP is just a waste of resource. Unused drive slots and ports on mobo are much less of a problem than dangling AC bricks and TB cables.

Well said .

(what Apple conceives) modern "pro workflow": independently modular components that work and power on its own, inter-connected via sub-optimal I/O that otherwise enables flexibility (eg. eGPU that can stay home while you take MBP out, or TB RAID enclosure that is not bound to this and only this tower)

The question is where Apple is positioning itself within this spectrum.

Apple has tried that approach for several years . That's if you consider limiting its hardware and leaving the rest to 3rd party suppliers as a viable concept .

My question is whether Apple understands that they are in no position to position themselfes in one way or another .
Every single Mac in their portfolio has been designed into a corner by now .

In 2006 they did the right thing by going mainstream and current with Intel - not entirely, mainly because Apple is too lazy to adapt GPU drivers - but it worked very well for many customers and Apple .

Today, it's mainstream or bust for the MP . There is virtually no use for a proprietary workstation running a proprietary OS , apart from said OS being much more convenient to use than Windows .
[doublepost=1535477279][/doublepost]

 

[ssgbryan]

Today, it's mainstream or bust for the MP . There is virtually no use for a proprietary workstation running a proprietary OS , apart from said OS being much more convenient to use than Windows .
[doublepost=1535477279][/doublepost]

Especially since they gutted the back end of their ecosystem. And the whole OS being much more convenient to use than Windows is a myth. People work in programs, not operating systems.

Currently, I would get a major performance boost if I ditched OSX and moved to Windows. I'd have access to more software and more hardware.

 

[goMac]

You do understand how much more you are adding to the TCO, do you not? Dumping the internal drives made the Trashcan even less competitive with it's counterparts from HP and Dell. In my case, it would have cost me an additional $1,200.

I'd rather have any space/bandwidth on storage spent on M.2 or PCIe.

You could put dozens of spinning drives on a Thunderbolt connection and not come close to saturating it. Putting drive bays in the case is sticking people with a legacy technology that don't necessarily need it. Back in the day of USB2, you had to put the drive inside the case to get full speeds from it. These days, that's not true. Spinning drives are multiple factors slower than even the slowest I/O connection on the box.

Would you rather have an extra PCIe slot or some drive bays? I'd take the PCIe slot.

(Apple is also never going to sell a hard drive only Mac Pro ever again, so if it's a price advantage you're looking for by skipping the SSD, we could probably just move straight past that.)