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

I would argue though that every additional PCIe slot - up to a point - will help maintain

That "up to a point" number is likely smaller than 4. Selecting 4 because the initial Mac Pro had 4 is largely form over function inference. Here is a thread with a survey ( which probably has some demographic sampling issues but it is better than "old one had for so that is the right number" hand waving.)

https://forums.macrumors.com/threads/whats-occupying-your-pci-e-slots.1614365/

1. There is are enough empty slots there to illustration that average is not going to be between 3.5-4. So it is lower than 4. .

2. A common occurrence is a "boot" GPU. A GPU card that is in the mix of some directly (or indirectly snarfed from Apple funded work) card that user. In other words there is a default display card. So pragmatically 4-1 = 3 . That slot isn't for 'random'/arbitrary cards. Helping to fill that is a role that Apple has to play a part in. if Apple doesn't diligently play that role then the overall product has a major problem .

3. Even in the older Mac Pros it is split 2 x16 and 2 x4 . So lots of folks got buy with just one x16 slot so x8-x16 needs in the old system. That wouldn't change for a new one.

4. Thunderbolt v3 largely covers the bandwidth and the latency of a x4 slot. If a x4 standard slot is "covered up" by TBv3 they user can still get to the bandwidth. So going back to the survey, there are more a few examples of USB 3.0 ( some 3.1 ) cards stuffed into a x4 slot. USB 3.2 , 3.4 , 4.0 , and probably 5.0 probably can all be covered by a TBv3 . The function ( not the form) of keeping up with the Jones in USB ports probably would be fine for the next 8-9 years with a TBv3 port.

If Apple puts 4-6 modern TYPE-A USB 3.1 gen 1 ports on the revised Mac Pro in 2019 that system would have have very good USB coverage forward (and back) for 8-9 service life of that machine. When the PCH chipset bumps on a future iteration to provisiong 4 USB 3.1 gen2 that wouldn't require much change to the design.

One x16 and one x4 slot would cover just everyone on that survey ( without getting into the weeds of folks who need a 'pure' Nvidia GPU line up in their Mac Pro ... i.e., GPU fanboy wars. ). At four slots already on the diminishing returns slide.


and even regain Mac market share , while the opposite will lose more customers and attract no new ones .
Hence I believe more PCIe slots - at least 4 - would significantly increase profit short, mid and certainly long term .

From the overall Mac market spective the Mac Pro even when has a healthy machine is probably in the 1-2% range. It will help smooth out the 'slop'/noise in the other vast majority of the Mac market ( units up/down on refresh timings , competitive changes , etc.) but it is highly unlikely to move the needle at all in a significant way. At best it will slightly wiggle less.

If the iMac and/or the 13" laptops all take 2% the Mac Pro isn't going to be able to bend that back up to maintain. The Mac Pro has been largely comatose over the last 8 years. Over that time Mac units have largely outpaced or match the PC market. ( not quirky corner Quarters but overall yearly trends). Mac Pro isn't some hidden super growth submarket that Apple has ignored. The "maintaine or get back to nominal growth" is largely about the Mac Pro submarket itself, not the Mac ecosystem. The Mac Pro could get back to trending water against overall Mac growth if Apple cleaned up their act in this space.


While many, maybe most users would not benefit from them ( right away ) , there are whole segments of Mac users who moved to other brands after the tcMP fiasko , and others who would never consider getting into Mac which has restricted expandibility/upgradability .

There is a whole segment of Windows users to left windows and came to Macs. There is bleed both ways. Only looking at one direction isn't going result in good market analysis fro the Mac Pro.

There are tons of folks who want an xMac and the Mac Pro isn't going to be an xMac.

Every day 90+% of the classic PC market doesn't buy a Mac. Coming up with folks who don't want to buy a Mac is like talking about how the sky is blue. It was blue yesterday. Is is blue today. It is going to be blue tomorrow.

Same goes for RAM and storage .

When has Apple been off of the RAM upgrade in the Mac Pro space???? They were replaceable on the 2013 model. Why wouldn't that be true on the next iteration. That's largely hysteria driven.

Storage. I think Apple is largely on track to equating that with PCI-e. again there is presence in the survey of more a few PCI-e cards to provision SSDs. There is a difference between standard PCI-e slot and a M.2 PCI-e slot. Apple putting deadicated SSD 1-2 PCI-e slots ( m.2 ) in would be a move in the right direction. The function is provision slots for internal SSDs. Saying that has to be in x4 standard PCI-e slot is a form issue.


'Extra' SATA cable flopping around inside the enclosure is different from there being a solution in the storage capacity space inside the machine. ( same issue just because old system had 4 spinner sleds doesn't necessarily mean going to have 4 spinners now. ). What Apple needs to provide is more internal storage capacity than the rest of the Mac line up. That is not necessarily "best NAS box" level of storage.
[doublepost=1533313223][/doublepost]
...
Sure. You are right. I thought that too... but marketing the computer as faster JUST because of a Tx and improved cooling would mean admitting that the other machines are drastically thermally throttled. I don't think that would be the strategy from Apple.

The T-x are in the "Pro" Mac products now because it is expensive. As the T-series evolves, it is extremely that it will be rolled out to all Macs. They will all have them. Right now Apple is in a catch-22. They are in so few systems they are probably a little too expensive for the entry level priced Macs. but they need volume to make them more affordable. ( which can't get until roll the T-series out to maximum volume.).

booting and security is a common feature that all Macs needs. That is a huge chunk of what the T-series is about.

It is highly unlikles that small volume run systems like the iMac Pro and Mac Pro are going to get relatively super low run T-series chips. What Apple needs is volume ranges in the 10's of millions ( or maybe multiples of millions) , not 10's of thousands. They can work out the the kinks with these higher margin/profit systems first before while ramp to that that kind of high volume.

The iMacs are probably next ( I"m not so sure about the Mini or any "more affordable' laptop that may come in the Fall).


Sell it with faster RAM, higher real world turbo speeds (better cooling), faster buses, better wifi, higher expanding options, more ports, future upgradeability at the same price or lower (exluding the screen, obviously) and I may consider the same hardware one year later. :)

More RAM, not faster RAM would be a differentiator. The iMac Pro just has 4 DIMM slots. If the Mac Pro had 8 DIMMs slots that would be a differentiator. Apple would have to get over their mantra of filling all DIMM slots in every configuration (and the all but one exception) . [ Or crank of the base price by putting in eight 8GB DIMMs as the baseline. Have to start at 64GB minimally] The folks who pragmatically want to run large RAM disks is something that the iMac Pro doesn't cover well.

The busses don't have to be faster, but using all (or most of ) them is an issue.

if more ports means "pour on TB ports like ketchup" that is a bad idea. Port placement as much as "more" might help. (e.g, two USB ports on the front of the system. Maybe a SD-Card port on the front. ). Outnumbering the iMac ports isn't so much an issue. That "sit and spin" feature of the Mac Pro 2013 wasn't well motivated. All the more so, if Apple goes back to deskside/underdesk default placement of Mac Pro.
 
  • Like
Reactions: barmann
Thanks for all your answers and the links, deconstruct60. :)

I must say I do agree with you on the Vega2 being sold later... after all these years I realised I lost the perception of what "interchangeable GPU" meant :D



Sure. You are right. I thought that too... but marketing the computer as faster JUST because of a Tx and improved cooling would mean admitting that the other machines are drastically thermally throttled. I don't think that would be the strategy from Apple.
[doublepost=1533306600][/doublepost]Sell it with faster RAM, higher real world turbo speeds (better cooling), faster buses, better wifi, higher expanding options, more ports, future upgradeability at the same price or lower (exluding the screen, obviously) and I may consider the same hardware one year later. :)

They wouldn't mention the cooling at all. They could also just ship them with the full -W chips instead of the slightly cut-down specialty ones the iMac Pro ships with.

But I don't think it matters either way. If you're buying an iMac Pro, you're buying an iMac Pro because it fits your use case and needs. If you aren't, it's because it doesn't. Presumably the Mac Pro is for that latter group, so some level of overlap doesn't really matter. Apple historically hasn't cared about cannibalizing its own sales as long as itself is doing the cannibalizing (witness no Microsoft-like attempt to protect its iPod business.)

Plus, they wouldn't be upfront about releasing a new Mac and uncharacteristically forward about timeframes if they were interested in creating, as some conspiracy theorists on this board and elsewhere have suggested, a failed product to justify killing the line entirely. There's no profit motive involved in that, not to mention how much it would erode the goodwill of pros they apparently are pivoting to keep or recapture.
 
2. A common occurrence is a "boot" GPU. A GPU card that is in the mix of some directly (or indirectly snarfed from Apple funded work) card that user. In other words there is a default display card. So pragmatically 4-1 = 3 . That slot isn't for 'random'/arbitrary cards. Helping to fill that is a role that Apple has to play a part in. if Apple doesn't diligently play that role then the overall product has a major problem .

The "boot" GPU problem would simply go away, if the Apple mMac Pro re-design were to include a PC-standard UEFI bios instead of their non-standard EFI scheme, which is only there for anti-Hackintosh purposes. Surely there is a better way for them to attempt to prevent the Hackintosh, while still retaining the ability to use standard and un-modified PC graphic cards (as the "boot" GPU)?
 
Last edited:
The "boot" GPU problem would simply go away, if the Apple mMac Pro re-design were to include a PC-standard UEFI bios instead of their non-standard EFI scheme, which is only there for anti-Hackintosh purposes.

Revisionist history hooey. Before the transition to x86, Macs were on OpenFirmware. At the time the rough equivalent on the x86 side was EFI. Apple jumped to EFI because they extremely likely didn't want to be trapped in the BIOS tar pit.

UEFI didn't stably finish until after Apple had already shipped the 2006-2007 sets of Mac products.

https://en.wikipedia.org/wiki/Unified_Extensible_Firmware_Interface#History

UEFI also went out of its way to loop back in BIOS with the compatibility mode. For first several years (2006-2009) most boards with UEFI hid it by default. UEFI was a mechanism in part to extent the life of BIOS. That isn't going to get high priority development resources at Apple.

There reason why the Windows side put tons of effort into dragging BIOS into the future was because of all the legacy infrastructure, cards, etc that was on BIOS. Well same thing is true now of Apple's EFI as the largely same excuse the mainstream vendors have draging along BIOS. ( there are more several year old Macs than 1-2 year old ones. ).

Yes along the way Apple did BootCamp and have their own kludge for BIOS? But folks are kidding themselves that Apple would not be more than happy to get past phantom MBR on drives and the patchwork of hacks to enable BIOS APIs.
Few on the mainstream side want to put BIOS to sleep. Intel announced they were kicking the can down the road to 2020.

https://www.anandtech.com/show/12068/intel-to-remove-bios-support-from-uefi-by-2020

EFI started in 2000. The handoff to UEFI foundation was in 2005 (and dragged on for several years before got to something that committee finally settled on ). 20 or 15 years and still BIOS is still hanging around.


Primarily an agenda against Hackintosh is a joke. People have to do some work to do a Hackintosh but Apple isn't dropping major DRM on stopping them. The mainstream market dragging their market deployment of UEFI and extensive retrofits back to BIOS are at least as responsible for the fork between Apple and the rest of the market anything Apple has done. The bigger the forces committed to forking the longer Apple is going to the fork they are on.


Surely there is a better way for them to attempt to prevent the Hackintosh, while still retaining the ability to use standard and un-modified PC graphic cards?


Apple has been incrementally closing the gap. As older Apple firmware drops into obsolesce and Apple incrementally folds in changes. Apple has enough to do eGPU at this point (without boot services ) which is market sufficient for them. IF Apple makes increments until 2021-22 then could merge in with the mainstream PC market when it has more firmly backseated BIOS. [ I haven't seen any other vendors besides Intel get on the "out by 2020 'bandwagon, so it is quite likely the foot dragging will continue of both sides for at 2-3 more years. ]
 
Apple has been incrementally closing the gap. As older Apple firmware drops into obsolesce and Apple incrementally folds in changes. Apple has enough to do eGPU at this point (without boot services ) which is market sufficient for them. IF Apple makes increments until 2021-22 then could merge in with the mainstream PC market when it has more firmly backseated BIOS. [ I haven't seen any other vendors besides Intel get on the "out by 2020 'bandwagon, so it is quite likely the foot dragging will continue of both sides for at 2-3 more years. ]
I'm not buying the whole "Apple needs to transition off EFI" argument. Apple has rarely, if ever, had a problem dropping older technology like a hot potato. I see no reason why they would hesitate today. Not that I am supporting the idea EFI is used to prevent hackintosh systems.
 
  • Like
Reactions: crjackson2134
Revisionist history hooey. Before the transition to x86, Macs were on OpenFirmware. At the time the rough equivalent on the x86 side was EFI. Apple jumped to EFI because they extremely likely didn't want to be trapped in the BIOS tar pit.

UEFI didn't stably finish until after Apple had already shipped the 2006-2007 sets of Mac products.

https://en.wikipedia.org/wiki/Unified_Extensible_Firmware_Interface#History

UEFI also went out of its way to loop back in BIOS with the compatibility mode. For first several years (2006-2009) most boards with UEFI hid it by default. UEFI was a mechanism in part to extent the life of BIOS. That isn't going to get high priority development resources at Apple.

Actually, if Apple wanted to go anti-Hackintosh, they had a better way to do it.

Open Firmware actually can work on Intel boards. I don't think anyone ships it for Intel boards, but Apple was considering it as an option.

So Apple had three options:
1) Ship with Open Firmware (which would preserve compatibility for existing Mac cards and other hardware)
2) Ship with traditional BIOS (which is what the developer Intel Mac kit did)
3) Ship with EFI (UEFI was not out yet)

Which would have been the worst for Hackintoshers? It wasn't EFI....
 
1. There is are enough empty slots there to illustration that average is not going to be between 3.5-4. So it is lower than 4. .
I don't think that for a professional workstation that it makes any sense to design for the "average" - it makes more sense to design for the 95% or 98% to satisfy the vast majority. (Especially the "power pros".)

Also, looking at that poll is misleading in that many of the cMP systems have a USB 3.0 card or a 10 GbE card - those should be unnecessary in the MP7,1.

Examine some real world cases. Dual double-width GPUs should be a requirement - and neither GPU should block a slot.

Many SAS and FC cards are x8 (96 Mb/sec is common for a dual port SAS card). FC could be eliminated by using real converged server-class NICs that can do FC with full offloading. What cards are needed for a high end audio workstation?

I'd put the minimum at something like:
  • x16 (double-wide)
  • x16 (double-wide)
  • x8 (x16 physical)
  • x4 (x16 physical)
 
(copy/paste from Zen thread)
Ryzen Threadripper 2950X - 16C/32T, 180W TDP, 3.5 base/4.4 Turbo, 899$
Ryzen Threadripper 2920X - 12C/24T, 180W TDP, 3.5 base/4.4 Turbo, 649$
(/copy/paste)
Not that Apple should build mMP on AMD (I wouldn't mind), but competition should give the fruit company a lot room to bargain with Intel.
 
Last edited:
(copy/paste from Zen thread)
Ryzen Threadripper 2950X - 16C/32T, 180W TDP, 3.5 base/4.4 Turbo, 899$
Ryzen Threadripper 2920X - 12C/24T, 180W TDP, 3.5 base/4.4 Turbo, 649$
(/copy/paste)
Not that Apple should build mMP on AMD (I wouldn't mind), but competition should give the fruit company a lot room to bargain with Intel.
More importantly - what Apple will do when AMD Rome comes out?

It will be faster than any Intel CPU, and anything they can offer. Would they switch to AMD?
 
  • Like
Reactions: singhs.apps
That "up to a point" number is likely smaller than 4. Selecting 4 because the initial Mac Pro had 4 is largely form over function inference. Here is a thread with a survey ( which probably has some demographic sampling issues but it is better than "old one had for so that is the right number" hand waving.)

https://forums.macrumors.com/threads/whats-occupying-your-pci-e-slots.1614365/

1. There is are enough empty slots there to illustration that average is not going to be between 3.5-4. So it is lower than 4. .

2. A common occurrence is a "boot" GPU. A GPU card that is in the mix of some directly (or indirectly snarfed from Apple funded work) card that user. In other words there is a default display card. So pragmatically 4-1 = 3 . That slot isn't for 'random'/arbitrary cards. Helping to fill that is a role that Apple has to play a part in. if Apple doesn't diligently play that role then the overall product has a major problem .

3. Even in the older Mac Pros it is split 2 x16 and 2 x4 . So lots of folks got buy with just one x16 slot so x8-x16 needs in the old system. That wouldn't change for a new one.

4. Thunderbolt v3 largely covers the bandwidth and the latency of a x4 slot. If a x4 standard slot is "covered up" by TBv3 they user can still get to the bandwidth. So going back to the survey, there are more a few examples of USB 3.0 ( some 3.1 ) cards stuffed into a x4 slot. USB 3.2 , 3.4 , 4.0 , and probably 5.0 probably can all be covered by a TBv3 . The function ( not the form) of keeping up with the Jones in USB ports probably would be fine for the next 8-9 years with a TBv3 port.

If Apple puts 4-6 modern TYPE-A USB 3.1 gen 1 ports on the revised Mac Pro in 2019 that system would have have very good USB coverage forward (and back) for 8-9 service life of that machine. When the PCH chipset bumps on a future iteration to provisiong 4 USB 3.1 gen2 that wouldn't require much change to the design.

One x16 and one x4 slot would cover just everyone on that survey ( without getting into the weeds of folks who need a 'pure' Nvidia GPU line up in their Mac Pro ... i.e., GPU fanboy wars. ). At four slots already on the diminishing returns slide.

Good points .

However - Apple is likely to release just one single basic design of the next MP , unlike their competitors .
Hence, they need to cover as much ground as possible, cater to as many user needs as possible in one single design .
In other words, a cMP 2.0 without the increasingly more important GPU support limitations .

As for PCIe slots, it doesn't matter how many there are for most prospective customers, but whether there are enough for the most demanding and least flexible customer .
Again the same is true for storage and memory, and again average needs don't matter when you consider market share and trickle down effect .

A lack of expandability and flexibility in a flagship model will lose bulk buyers instantly, especially when the rest of the offerings follow suit .
Eventually, Macs will become even more of a fashion item than they already are , without a foothold in any industry since noone is being trained on OSX machines any longer .




There is a whole segment of Windows users to left windows and came to Macs. There is bleed both ways. Only looking at one direction isn't going result in good market analysis fro the Mac Pro.

I doubt that .
I have no proof or numbers, but I very much doubt that there is a significant number of people moving to OSX .
Hardware and software support by and for Apple is as low as it's ever been ; casual users might go back and forth, major outfits have no reason to favor Apple over more innovative and reliable partners, for everyone else only the cost and effort of leaving OSX behind is a factor .



There are tons of folks who want an xMac and the Mac Pro isn't going to be an xMac.

Every day 90+% of the classic PC market doesn't buy a Mac. Coming up with folks who don't want to buy a Mac is like talking about how the sky is blue. It was blue yesterday. Is is blue today. It is going to be blue tomorrow.

Trouble is, that sky is getting bluer by the day .



When has Apple been off of the RAM upgrade in the Mac Pro space???? They were replaceable on the 2013 model. Why wouldn't that be true on the next iteration. That's largely hysteria driven.

Storage. I think Apple is largely on track to equating that with PCI-e. again there is presence in the survey of more a few PCI-e cards to provision SSDs. There is a difference between standard PCI-e slot and a M.2 PCI-e slot. Apple putting deadicated SSD 1-2 PCI-e slots ( m.2 ) in would be a move in the right direction. The function is provision slots for internal SSDs. Saying that has to be in x4 standard PCI-e slot is a form issue.

Two things : cost and backwards compatibility .

RAM upgrades that become affordable years after the initial release of a computer are a joke .
Apple has not ever build a computer or anything else to be cutting edge or be aimed at the future ( insert trash can joke here ) , why start now ?

Storage . By all means, prepare it for m.2 , make it compatible !
As in optional ; right now, SATA SSDs are still significantly cheaper than m.2 variants, and there are tons of them around people already have .

Are we supposed to buy all new stuff that is faster, but beneficial for just a fraction of uses at this point in time, or spend a ton on external enclosures to keep using perfectly fine drives ?

I'm sick and tired of technology that exists for its own sake, which dangles a carrot in front of my nose and inevitably turns out to be a chimera .
And which gets decided upon in mid level management meetings .
 
One thing I’ve been thinking about - in Apple’s discussions about the new Mac Pro the language they have been using is not “upgradeable” or “expandable” but “modular”.

There seems to be an assumption that a new Mac Pro will be a reversion back to the previous Mac Pro style - tower-type chassis with a bunch of slots for RAM, cards, sockets for CPUs and fans, etc. I’m not sure that’s what Apple intends to offer. Presumably they could easily ship such a Mac, but instead they are taking two years over it. The fact that they have made such a song and dance about going deep into understanding the needs of “Pro Workflow” perhaps suggests that they don’t think that’s the answer.

It also seems fairly clear that “upgradeable Macs” is not an design principle or commercial principle that Apple generally holds, as they have progressively removed such capability across the whole range. So I don’t think it makes sense that Apple would undermine that with a new Mac Pro.

I think there is a fair chance that the new Mac Pro that Apple eventually ships is something that is modular in terms of hardware accessories that allow it to be optimised for different pro workloads, but is not upgradeable or expandable in the way people want or expect.

Just my views.
 
One thing I’ve been thinking about - in Apple’s discussions about the new Mac Pro the language they have been using is not “upgradeable” or “expandable” but “modular”.

Thunderbolt does not have, and can not have the throughput of a PCI slot. Apple has talked about the next Mac Pro being optimised for throughput. It's just as useful a reading of the chicken entrails.
 
  • Like
Reactions: ssgbryan
Few know what Apple will release with the next Mac Pro, everyone else is merely speculating based on their ideas of what they'd like it to be.

IMO the problem isn't what the next Mac Pro will or will not be but the lack of interest Apple has demonstrated for the Mac Pro. If I were a professional who relied on the Mac Pro for my livelihood I'd be working on my migration plan away from the platform. The stagnation and lack of communication are deal killers.
 
  • Like
Reactions: Aldaris and Biped
Thunderbolt does not have, and can not have the throughput of a PCI slot. Apple has talked about the next Mac Pro being optimised for throughput. It's just as useful a reading of the chicken entrails.
When and where have they talked about the next Mac Pro being optimised for throughput?

In my experience Apple is extremely deliberate in their choice of language.
 
IMO the problem isn't what the next Mac Pro will or will not be but the lack of interest Apple has demonstrated for the Mac Pro. If I were a professional who relied on the Mac Pro for my livelihood I'd be working on my migration plan away from the platform. The stagnation and lack of communication are deal killers.

After thinking, reading - and whining - about this topic for 6 years(!) now, I have come to the conclusion that the mmMP (aka the "mythical modular MacPro") won't be like 99% of the folks here (and elsewhere) want it to be. It is not only what you outlined above, but also look at the recent development (aka iMac "Pro" or macBook "Pro"). For me, Apple completely has lost track of what about a sizable share of pro(sumer) users actually want. Instead, even the machines formerly known as "pro" machines now have become more like a lifestyle item, and "form follows function" essentially has become "function follows form", thus violating one of the most basic and important design principles ever. Earth to Jony, anybody out there? Butterfly keyboard is a good example, as well as the thermal restrictions of both the nMP and the iMac "Pro" (edit: and wait, same seems to apply to the lastest MBP as well). And since they now seem to repeat their failures over and over again (i.e. iMac Pro thermals, and does anyone remember the cube?) why not do it again?

So rather than completly admitting their failure, I guess Apple rather wants to "innovate" (their asses) for the sake of innovation rather than doing the obvious. I completely have lost faith in them at this point. Apart from the fact that the mmMP will be a quite expensive thing no matter what they do. Of course I might be wrong, but I'd say: don't hold your breath.

However, I still want to stay on the platform, because for me it is not the hardware, but actually macOS that keeps me there. So now, I decided to do exactly the same what I did five years ago, i.e. show Apple my middle finger, and build my next hackintosh. There seem to be some very good builds around the X299 platform, even including TB3 support.

So at least for me, the wait is now over!
 
Last edited:
If I were a professional who relied on the Mac Pro for my livelihood I'd be working on my migration plan away from the platform. The stagnation and lack of communication are deal killers.

Apple has demonstrated an inability to navigate the future outside of platforms in which it controls every element of the technologies. They got it wrong on the speed of USB-C adoption, they got it wrong on Intel's processor progress, the future direction (and who would produce the goods in terms of performance/power) of GPUs - the irony is that it's Apple projecting its wish fulfilment upon the industry at large that keeps biting them. Unless they can accept that, and move to hardware that, as its core design philosophy, is a skeleton that third party hardware fleshes out, then the Mac Pro has no place, answers no questions, and serves no purpose, other than being "the fastest Mac".
[doublepost=1533482740][/doublepost]
When and where have they talked about the next Mac Pro being optimised for throughput?

I think it was Shiller on a talk show live or something similar.
 
the irony is that it's Apple projecting its wish fulfilment upon the industry at large that keeps biting them. Unless they can accept that, and move to hardware that, as its core design philosophy, is a skeleton that third party hardware fleshes out, then the Mac Pro has no place, answers no questions, and serves no purpose, other than being "the fastest Mac".

Well put. I don't think a Mac Pro makes any damned sense with the presence of an iMac Pro unless it is that skeleton you speck of. Now if Apple sees it that way, who knows.
 
However, I still want to stay on the platform, because for me it is not the hardware, but actually macOS that keeps me there. So now, I decided to do exactly the same what I did five years ago, i.e. show Apple my middle finger, and build my next hackintosh. There seem to be some very good builds around the X299 platform, even including TB3 support.
I understand a preference for a particular OS but shouldn't the decision be based on the applications that run on the OS?
 
  • Like
Reactions: ssgbryan
I understand a preference for a particular OS but shouldn't the decision be based on the applications that run on the OS?

Yes, but when general app support is more or less on par? I am doing mainly frontend web development, and for this task macOS is the most perfect OS you can have (imho).

Of course I could do my work on windows as well, as most of my apps are available there as well. But when I solely look at the OS itself, macOS still wins hands down compared to win 10 for various reasons, one of them macOS being POSIX compliant (makes maintaining a local dev server more easy), another one is the general UI/UX and of course while I do not trust Apple that much, I still trust them more than MS when it comes down to things like privacy. Again YMMV, but for me this decision is a no brainer - so much that I am willing to take the additional hassle of building and maintaining a hackintosh. I have been doing this for five years now anyways, so yeah let's go on with this ;)
 
I mis-spoke earlier when I said "single vs dual GPU is a configuration option"... What I was trying to say is that I believe Apple will offer both options. I think we'll see two x16 slots (or some equivalent), but we may very well (probably) see something that restricts GPU choice (no NVidia).

One possibility would be slots that were electrically PCIe 3.0 x16, but not in the standard PC form factor (probably for the card - the connector itself may well be standard). If they are injecting the video into the Thunderbolt 3 ports, which seems likely at least for the primary card, they don't even need the rear connector panel. I wouldn't be at all surprised if the video is a card (or cards) with no rear access that actually sits parallel to the motherboard. It would have some extra connector that carried the video signal back to the motherboard (or an I/O board) to go to the TB3 ports. Before saying "no way", name any Mac since the cheesegrater that has a video port located on the video card (a bunch of iMacs have had replaceable video cards, as does the cylinder Mac Pro, but the video is carried back to the motherboard)?

Even if the card is a size that a PC-type card will fit, possibly some "short" form factor from mini-ITX PCs, the Thunderbolt injection might be nonstandard enough to prevent regular cards from working. I can see the machine refusing to boot if it doesn't detect a GPU injecting video onto the Thunderbolt bus.

Never underestimate Apple's ability to do something weird on purpose just to keep people from using commodity PC parts, especially if there is something (NVidia cards) that they are trying to keep people away from... It wouldn't shock me if the card was a standard form factor and even had access to the back panel, but Apple screwed with something like the auxiliary power connectors (which are a kludge on normal PCs) or the EFI implementation.

I do expect at least one PCIe slot for audio interfaces and the like that is relatively standard (but maybe x4, and probably physically fairly short). Many other interfaces will be taken care of, so there is less need for PCIe cards than on the cheesegrater. Many of the cards people run on cheesegraters (as has been mentioned) are to add modern ports to a 2006 design. The new one will have 10 GBe and plenty of Thunderbolt 3 ports that break out to just about everything else.

I strongly expect standard RAM (most likely 8 slots), although accessibility is an open question. They'll have some way of adding internal storage, but almost certainly not spinning disks, and probably no SATA. It may be standard NVMe drives, or some Apple solution.

Dual processor support is a maybe, because it means Xeon-SP... We won't see a quad-core CPU in a low-end model, and probably not a 6-core, either. They don't want to release a Mac Pro that can be beaten by their own laptop (and any quad-core is at risk from the 6-core MBP in some tests!). They will probably steer clear of anything with 6 cores, too - unless they leave the iMac at 4 by either not updating it or using low end processors (any of the high-clocked 6 core i7s will beat the 6 core Xeon W models) - this only gets worse when 8-core desktops appear!
 
Last edited:
Yes, but when general app support is more or less on par? I am doing mainly frontend web development, and for this task macOS is the most perfect OS you can have (imho).
Unless you're using the OS, and only the OS, for your web development the OS is not as important.

Of course I could do my work on windows as well, as most of my apps are available there as well. But when I solely look at the OS itself, macOS still wins hands down compared to win 10 for various reasons, one of them macOS being POSIX compliant (makes maintaining a local dev server more easy), another one is the general UI/UX and of course while I do not trust Apple that much, I still trust them more than MS when it comes down to things like privacy. Again YMMV, but for me this decision is a no brainer - so much that I am willing to take the additional hassle of building and maintaining a hackintosh. I have been doing this for five years now anyways, so yeah let's go on with this ;)
Normally I would say use what you prefer. However with Apple's distaste for the Macintosh I think it's reasonable to look at alternatives. One thing to keep an eye on is the Windows Subsystem for Linux capability of Windows 10. It's still young but might be something to consider if POSIX capability is a must have.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.