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

[deconstruct60]

....
The iPhone (it's software and AI development) is the reason for Apple to Offer again a Mac Pro, its no secret from Ive to Cook they wanted to kill it in basis to its development take R&D resources from other priority projects,

It isn't really so much as "take resources" as Apple's throttle on expanding available resources. Steve Jobs put a "behave like a small company" limiters on saying no to most projects. It is a dual edge sword. For example, if you run all of your Wi-Fi subsystems through a single "wi-fi subsystem problem' group then knowledge picked up on product A can be more easily applied to product B since is it the same relatively limited group the institutional knowledge gets applied. You don't have folks re-inventing the wheel.

The back edge of the sword is that at some point that functional team is only going to have so much bandwidth. The other danger over the long term is group-think. I think Apple tries to mitigate the group-think by dropping products in mature, shrinking markets and pushing groups into newer areas with higher potential growth and new problems.

and they then didnt foresee to need a powerful workstation for VR/AI, glad VR/AI grown we will have a mMP again but just the MP Apple needs for its iOS developers, so those expecting 4 GPU, std PCIe better dont embrace big expectations.

Actually AI and lust for 4 GPUs set-ups more aligns with the MP 2013's standard configuration mandated dual GPUs than the narrow demand for duals that Apple pointed to as one of the limitations of the MP 2013's design. Significant demand for more grunt in a single than spreading things out.

Apple's tag of AR will reach widespread market deployment first before VR does is still firmly on track. VR isn't rapidly scaling across the iOS market right now (or will in next year or so). Similarly having the local inferencer is one of the critical pieces of AI. ( if can't use what you have learned then it is going to be hard to add value to end users. )


Shouldn't confuse latest tech porn hype trend with real demand in the workstation market. There aren't that tightly coupled. Especially among the "I want to use internal components I already bought or want to buy " crowd.

There are much bigger "didn't foresee" that boat anchored the MP 2013. Apple somewhat walking away from OpenCL (and heading off to Metal ) I don't think was weaved into that long term plan for the Mac Pro at all. The Mac Pro had a large bet on OpenCL and that didn't line up with what actually happened. There were several players outside of Apple's control that heavily influenced that, but none the less the MP got caught in that major shift. If the compute GPU software base didn't expand at a significant enough rate the MP would have issues.... and that happened.

GPU evolution moving to newer process to get to better managed thermals crapped out. Apple has been relatively immune because their GPUs are relatively small and they can charge a higher than norm price for their SoC chips but GPU market (and AMD in particular) hit a wall. "Big" GPU just went hotter max TDP over last 4 years which wasn't what Apple (or GPU vendors were expecting. Nvidia adapted a bit better, but it wasn't what they were talking 4-6 years ago. )

At least Apple build a tcMP prototype in 2016 loaded with dual RX480-family GPU and 6 TB3+4(10) USB3 ports, there where leaked evidence of that, and someone at NAB watched this prorotype and leaked it, for some reason Apple decided not to sell it, I dont buy the "blame the thermal corner" as reason for that, since there where N turnarrounds to raise the tcMP TDP to 600W enough for dual RX580(full clock) or dual Vera64(underclocked), and a typical Xeon E5v4

The problem is that the 480 (and 580) are roughly the evolutionary equivalent of the AMD Pitcairn offerings. (mid range). Apple could have upgraded roughly the same Mac Pro into new D300 replacements that did better than the old D700 but they'd have nothing to equally move the D700's. There would have been some upside in just simply doing something, but also would have put a bigger spotlight on the limitation. ( Eventually they had to spotlight it anyway so it would have been a better "stop gap" for this gap time while they restart from scratch. But if that would have slowed up the iMac Pro that would have been a problem. )

Vega weren't a solution because they are quite late and limited even then. Nvidia had whatever custom design roadblock that existed in 2012-2013 ( probably made worse by Nvidia roaming around threatening to sue mobile GPU makers ).

Hand waving and cranking to 600W wouldn't work. While Apple could have done better job using the a higher fraction of the single fan's diameter to move air ( and could have widened he diameter a bit) while at the same decibel levels, but at some point get into a slippery slope where more than one fan is needed and at that point baseline presumptions are off. iMac Pro has two fans.

The thermal corner wasn't the sole issue. Another stated one was the workload breadth limitations on the dual GPUs. Once that is on the questionable block some elements of the baseline MP 2013 design assumptions don't fit.

While Apple didn't mention it 6 TBv3 ports don't work all that great either. It was questionable for the Mac Pro 2013 design. It is vastly more questionable now with higher bandwidth (and utility) for TBv3 in a context where likely using discrete 3rd party monitors. The iMac Pro doesn't do six. There is x4 PCI-e (besides the x16 for 2nd big bandwidth use ) left on the floor in the iMac Pro. Prior to the arrival of Intel W class processor (and chipset) 6 port is just slavishly copying the old design. It is goofy.

 

[deconstruct60]

...

They literally didn't start on it until after the roundtable, and they had to put together a new team, along with getting the iMac Pro out the door.

I don't think they had solidified anything but I suspect they had started laying the ground work (gathering market information as to who and what problem demographics were).

"... enables us to do these quick, regular updates and keep it current and keep it state of the art, and also allow a little more in terms of adaptability to the different needs of the different pro customers.

Lance Ulanoff (Mashable): I’m just curious, at what point did you realize that? ....
But at what moment in the product cycle did you think: ‘Oh… This is maybe not the end all, be all.’ Did that happen six months ago? Where did you get the telemetry that told you that?

Craig Federighi: I’d say longer than six months ago. But I think we designed ourselves into a bit of a thermal corner, if you will. ... "

https://techcrunch.com/2017/04/06/t...-john-ternus-on-the-state-of-apples-pro-macs/

The iMac Pro pretty much lays approximately the same thermal corner. it is two fans and largely decoupled CPU and GPU radiators, but approximately the same top end limit. 300W GPUs just aren't practical with the iMac Pro.

The iMac Pro's adaptability is even more narrow in some ways than the MP 2013 design. So it hardly could have been the solution to "regular updates and adaptibiility" breadth issues.

People here underestimate how long it takes to do a product when you're starting from scratch and the team already had another machine to get out the door.

Yes. "But Dell and HP got something out the door". isn't illustrative of the issue. Dell and HP had either pipelined teams working to different release points or different start points. The "normal cycle" for Xeon E5 (and now Xeon W) class tick-tock cycle completion was 1.5-2 years. So a new baseline design wasn't needed every year and if have a team start 2.2-2.3 years in advance you'd have something at the end of that 2 year cycle.

Hurry up and do it quick relatively smacked of the MP 2013 design in that the D700 GPUs really didn't fit. That the design got solidified before they had firm specs and the whole thing didn't work on long, peak job loads because nobody tested it.

We're probably just getting to the point where they might be doing DVT runs, but even then those aren't usually super leaky.

It isn't that it is DVT test runs. It is DVT test runs with a very small, very narrow set of contractors. The iPhone leaks like a sieve because they are trying to get several thousand people to keep a secret spread over 4-5 layers of sub, sub-sub , etc contractors. Eventually you have 2-3 layers that don't even deal with Apple directly so they could give a flip if Apple gets mad and attempts to track them down. There is a feeling they can just get lost in the mob.

The Mac Pro doesn't need massive logistics or significantly large DVT production runs to prove it scales up massively.

Apple really was caught with their pants down, and they had that roundtable as soon as they realized it. There is no master plan, no hidden conspiracy. They were starting with absolutely nothing, and they had absolutely nothing when Schiller was talking. The whole thing is a disaster.

I think hiccups with Vega and Intel W availability were an issue. That session wasn't solely about the Mac Pro. It was also that they were doing a new iMac like solution but it really wasn't going to be ready. The other big issue was that the MP 2009 were going to be dropping on to the vintage. Throw that on top of the press running "no Mac Pro for 1000+ days" (and the other Mac comatose product debacles ... Mac Mini not really all that far behind at that point and blown past that mark now. ) There was a baseline problem of does anyone even work on Macs anymore issue. That why they held the pow-wow in the "Mac design machine shop". There was an aspect of 'we are busy grinding and milling away at something" flavor to the talk.

The iMac Pro was going to be late covering the MP 2013 model. The needed to drop prices to better "stop gap" until they could get the iMac Pro out the door. ( which similar to the MP in 2013 somewhat slid into 2018 for the >10 core options. )

They were, they aren't now. I think they very quickly got feedback from MVP customers they privately demoed the iMac Pro to that it would NOT work was a Mac Pro replacement. THEN they had the fire drill roundtable once they realized the iMac Pro couldn't replace the Mac Pro.

Really? It if far more likely the bulk of the "MVP" customers who are most bent out of shape by the iMac Pro are the largely the same as the ones that were bent out of shape about the Mac Pro 2013 in mid-late 2014. That group shouldn't be a "recent news flash" to Apple.


Certainly Apple probably agitated a small fraction of the folks who were happy with the MP2013 approach by coupling the LCD panel and taking away easy access to RAM upgrades, but that's is probably far smaller than the group who didn't like the MP 2013 in the first place. What Apple is doing is adding that relatively small group to the other relatively (to the larger trends) group to form a perhaps sustainable group for another stab at the Mac Pro.

 

[goMac]

It isn't really so much as "take resources" as Apple's throttle on expanding available resources. Steve Jobs put a "behave like a small company" limiters on saying no to most projects. It is a dual edge sword. For example, if you run all of your Wi-Fi subsystems through a single "wi-fi subsystem problem' group then knowledge picked up on product A can be more easily applied to product B since is it the same relatively limited group the institutional knowledge gets applied. You don't have folks re-inventing the wheel.

Before 2012 the Mac Pro actually had it's own dedicated team, but once Apple started re-organizing, they integrated the Mac teams together.

So it's true that the Macs generally compete for resources. It's very likely the people who were working on the Mac Pro had to spend time on the iMac Pro (and possibly the MacBook Pro). But you aren't going to have the Mac Pro competing with something like iCloud or iPhone.

Another thing not to discount is that it's very likely a lot of the original Mac Pro team is gone. Not reassigned, just gone. Not only is Apple starting over on the Mac Pro, they're starting over on the Mac Pro team.

It's why I tend to doubt rumors of things like AMD CPUs. They're really in a bind and I can't see them doing extra tasks.
[doublepost=1522094045][/doublepost]

Craig Federighi: I’d say longer than six months ago. But I think we designed ourselves into a bit of a thermal corner, if you will. ... "

So here's the very last thing I heard... There was a time period in which the iMac Pro was being considered for replacing the Mac Pro. It may be the six month window Craig was referring to.

But... Apple did not decide to do a new Mac Pro until right before that meeting. When they realized how long that would take, they called that meeting to explain the time gap until the next Mac Pro.

I know for sure when that meeting was held there were no ideas for the next Mac Pro on the table besides "we need one." There was no work done. That announcement was even a surprise to some people inside Apple's hardware team.

I'm not sure if people knew exactly how WTF'y the whole thing is that it would make them feel any better about buying a Mac. It certainly is not an intentional set of actions leading Apple to some master plan.

Yes. "But Dell and HP got something out the door". isn't illustrative of the issue. Dell and HP had either pipelined teams working to different release points or different start points. The "normal cycle" for Xeon E5 (and now Xeon W) class tick-tock cycle completion was 1.5-2 years. So a new baseline design wasn't needed every year and if have a team start 2.2-2.3 years in advance you'd have something at the end of that 2 year cycle.

Right. Dell and HP have existing teams, Apple is starting from scratch.

Really? It if far more likely the bulk of the "MVP" customers who are most bent out of shape by the iMac Pro are the largely the same as the ones that were bent out of shape about the Mac Pro 2013 in mid-late 2014. That group shouldn't be a "recent news flash" to Apple.

I'll just quote myself...

I'm not sure if people knew exactly how WTF'y the whole thing is that it would make them feel any better about buying a Mac.

 

[deconstruct60]

So here's the very last thing I heard... There was a time period in which the iMac Pro was being considered for replacing the Mac Pro. It may be the six month window Craig was referring to.

....

I know for sure when that meeting was held there were no ideas for the next Mac Pro on the table besides "we need one." There was no work done. That announcement was even a surprise to some people inside Apple's hardware team.

If "we need one" has some concrete parameters to it then there was some work on collecting from core requirements. It is initial ground work, but it is something.

I'm not sure if people knew exactly how WTF'y the whole thing is that it would make them feel any better about buying a Mac. It certainly is not an intentional set of actions leading Apple to some master plan.

"Leave a hole and do almost nothing on this product" is a plan. It is consists of doing nothing in that particular area but it is a choice of action. I'm not quite buying the 'catch with pants down" characterization of this.

Right. Dell and HP have existing teams, Apple is starting from scratch.

This too is seems to be somewhat of stretch characterization. They have the working iMac Pro motherboard. So they have dealt with Xeon W and Vega , 10GbE , and 4 ports of Thunderbolt. Mutating that into a bigger boards and the GPU based off of a derivative of AMD reference board is work but if have some above average talent that did the iMac Pro work that shouldn't be a huge show stopper. ( unless pulled that iMac Pro logic board team from the iMac team. That isn't scratch. That is far more so just under resourced; to few folks spread waaaaasaaaaay too thin. )

The iMac Pro could pick up speed bump, socket compatible CPU and BGA compatible GPU updates later that the Mac Pro could use ( it is pretty close to the 2009 -> 2010/2012 work). So if merge most of the large components one team could handle both and the team would already have experience doing the iMac Pro. In terms of getting the share components to boot and do basic macOS services should be pragmatically almost identically the same.

Apple should have some kind of team working on Thunderbolt peripherals and eGPU issues. So the drivers and software dealing with std PCI-e slot just need to liaison with the Mac Pro team. That shouldn't be "starting from scratch" because all the TBv3 capable Macs have to deal with it.

There are some some differences but if Apple stuck with standards: std PCI-e slot , standard M.2 slot and some qualified tested 3rd party SSDs , and perhaps yank a couple drive sleds from the MP 2010 models. Starting from scratch notion is kind of week. This is stuff from AMD and Intel reference board designs, it should not be all that hard to copy into a Apple design compatible form.

There are difference though so it isn't a wham-bam 2 month job. The iMac Pro only has to deal with two fans. The thermal environment in a Mac Pro would be a bit different: perhaps 3-4 thermal zones. The T2 chip apparently does some of the Fan control and if it is capped out at 2 then either have to work around or get an update. The cooling systems are going to be different. Probably need a different RF path out of the case than the iMac Pro uses. Should have 8 DIMM slots instead of just 4. The layout flow of the iMac Pro needs to be redone.

The industrial design team is probably also going to be choke point.

I'll just quote myself...

So it is largely just the same group. ;-) Honestly this looks like a marketing (as in real collection about customer usage/demographicss/needs ; not 'sales spin' which is really sales. ) and product management failure to push back about being under resourced. The "master plan" smells like one optimized to put the max into the Scrooge McDuck money pit that drives up the stock value so the middle-upper management can cash in their discounted stock to "keep up with the Joneses" in the crazy expensive Silicon Valley cost of living/housing market. It is a bad long term plan for the Mac but hardly a "WTF" moment in Silicon Valley these days.... all too often a common game plan.

 

[CWallace]

Apple likes their products to do something competitor products either cannot do or cannot do as well. TouchID. FaceID. The A-series CPU/GPU. The M-series MPC. Retina Displays. The W-Series audio chip in the Air Pods. All of these are things competitors do not have, or their equivalents are not as effective / seamless / integrated.

Outside of the design, the only difference between a Mac and a PC is the Operating System. Hardware-wise, there is nothing to distinguish them apart. The 2009-2012 Mac Pro was no different from an HP or Dell tower workstation on a component basis. As such, I understand why many just want Quanta Computers or whomever makes HP's workstations to just slap an Apple Logo on the front of a Z8 instead of the HP logo, install macOS instead of Windows or Linux, and call it good. And Apple could have done that back in April if that was what they wanted.

But Apple seems to be trying to bring the same "innovation" to the Mac that they have to their other products. The T1 chip on the MacBook Pro with TouchBar and the T2 on the iMac Pro. The Touch Bar itself. Yes, I know many on this forum - and especially in this thread - don't give two s**ts about any of that and actively resent having to pay the premium Apple charges for those features.

Apple tried to "innovate" with the Late 2013 Mac Pro by moving the expansion from internal slots to external and pinning their hopes on software being developed to leverage the "compute GPU"'s power. TB2 didn't have the bandwidth to replace every PCI component, however, and most software did not leverage that second GPU's power. So even though Apple could have updated it with slightly better CPUs and GPUs, it would not have really affected sales because the original model either worked for you or it did not and slightly better CPUs and GPUs would not have changed the "nay" to a "yea".

So that Apple has not just re-branded an HP workstation means that they're trying to "innovate" again. They do seem to have taken the lessons learned from the 2013 model to heart with the words "modular", "expandable" and "updatable" added to the lexicon.

I know many are just going to assume Apple will f**k it up again like they did in 2013 so just give us the re-branded PC. And maybe they will and this will resolutely kill the high-end computing market for Apple.

But maybe they won't and what we get will have been worth the wait.

 

[goMac]

This too is seems to be somewhat of stretch characterization. They have the working iMac Pro motherboard. So they have dealt with Xeon W and Vega , 10GbE , and 4 ports of Thunderbolt. Mutating that into a bigger boards and the GPU based off of a derivative of AMD reference board is work but if have some above average talent that did the iMac Pro work that shouldn't be a huge show stopper. ( unless pulled that iMac Pro logic board team from the iMac team. That isn't scratch. That is far more so just under resourced; to few folks spread waaaaasaaaaay too thin. )

I tend to think any Mac Pro design will be based on the iMac Pro. We'll see what the latest and greatest Intel and AMD have at release time is, they might roll updated components.

I honestly think most their time being spent right now is WTFs spent on things like GPU upgradability. The iMac Pro doesn't bring the upgradability they want to have, and it doesn't solve the problems they want to solve. I don't doubt that "cartridge style" GPUs are an option, but I'm sure they're having debates on if that is something that would even work for customers.

With stuff like neural networks they might already know that consumer cards aren't going to cut it, especially after all the complaining about the 2013. If I were in their shoes, I would strongly be considering finding a way to make a PCIe slot work, so that way Apple doesn't have to sell their own Radeon Pros or Tesla cards. A proprietary slot will restrict the markets Apple can sell into severely and I think they'll try to avoid that. Apple won't directly sell Nvidia cards but they also know their customers want to be able to upgrade to one.

Yeah, eGPU could maybe solve that but I don't think that is where Apple wants to go with a desktop.

 

[deconstruct60]

...

I honestly think most their time being spent right now is WTFs spent on things like GPU upgradability. The iMac Pro doesn't bring the upgradability they want to have, and it doesn't solve the problems they want to solve. I don't doubt that "cartridge style" GPUs are an option, but I'm sure they're having debates on if that is something that would even work for customers.

That's pretty sad. Take the reference card. Strip off the external connectors for video out. Take those output lanes and route them back into a connector on the motherboard. Major issue large done. They could iterate on what the cooling system was but this really isn't a Moon Mission. Apple used some MXM like cards in the iMac several iterations back; they have used the core notion here before.

I think they need to take some time. They should pick a new connector system that they can use for several years ( a new standard slot). So will want something that can deal in the future with PCI-e v4 , a flexible range of power if power is part of the mix, displayPort 1.5-1.6 ( targets for HDR 8+K screens ). However, it doesn't have to be module they could re-use back into the iMac Pro. (that's a bit too Scrooge McDuck cost savings. )

With stuff like neural networks they might already know that consumer cards aren't going to cut it, especially after all the complaining about the 2013. If I were in their shoes, I would strongly be considering finding a way to make a PCIe slot work, so that way Apple doesn't have to sell their own Radeon Pros or Tesla cards.

standard for the secondary x16 slot. Sure. They should. Once again it seems to be a major flow to tightly coupled the 2nd slot with the primary display GPU slot. They have two different requirement sets.

They can easily come up with a design that works for their primary slot that will also work in a standard slot as primarily a "Compute GPU" (no display output needed). But not vice versa, standard GPUs don't fit in primary display slot. Apple shouldn't be trying to make "Compute GPU". There isn't going to be enough scale because not going to bundle them across all systems.

A proprietary slot will restrict the markets Apple can sell into severely and I think they'll try to avoid that. Apple won't directly sell Nvidia cards but they also know their customers want to be able to upgrade to one.

What other market is Apple going to sell into other than Macs? Apple isn't going into the general "add in card" GPU market any more than they are going into the "add in" SSD market. That isn't their customer base so why bother?

These are primarily bundled GPU cards. The price is built into the system so "price anchor" issues really shouldn't be a major problem. Upgrade costs, Apple shouldn't try to match Pro card prices. Higher than consumer will work with a reasonable subset of pros who are actually making money. Many of the load moaning , price anchored complainers are going to buy used/pulled anyway. They grumble at everything.... including new Mac Pro that Apple sells (tend to buy those used too. ). If Apple isn't selling enough standard configs GPUs to pay for the cards development costs then should just quit the whole thing. If have the card development costs covered then the upgrade market is just 'gravy'; stop being so greedy.

Yeah, eGPU could maybe solve that but I don't think that is where Apple wants to go with a desktop.

The system issue though is that Apple is going to have to deal with eGPUs coming from the rest of the Mac ecosystem other than the Mac Pro. It still means software and drivers support work they'll have to do. There seems to be a very flawed notion in Apple lately that if they can push it out of the standard configs sold then Apple can just wash they hands of the complexity of what folks are going to add back in via Thunderbolt ( or USB). Kick out all of the hard drives so APFS can just largely punt on those. Errr, that really isn't going work.

eGPU should be where Apple does go with the desktop ( because the MBP , iMac , and a non comatose Mac Mini ) will be on the desktop. The Mac Pro should minimize desktop footprint by being remote from the desktp ( side / under / mounted in a box / etc. ) and connect back to the desktop resident monitor by a cable.

I suspect Apple doesn't want to deal with the drama of 2-3 3rd party cards but containing a single one shouldn't be that hard of stretch.

 

[goMac]

These are primarily bundled GPU cards. The price is built into the system so "price anchor" issues really shouldn't be a major problem. Upgrade costs, Apple shouldn't try to match Pro card prices. Higher than consumer will work with a reasonable subset of pros who are actually making money. Many of the load moaning , price anchored complainers are going to buy used/pulled anyway. They grumble at everything.... including new Mac Pro that Apple sells (tend to buy those used too. ). If Apple isn't selling enough standard configs GPUs to pay for the cards development costs then should just quit the whole thing. If have the card development costs covered then the upgrade market is just 'gravy'; stop being so greedy.

I'm 99% sure Apple has always been losing money on GPU upgrades or breaking even.

They're limited runs of reference designs (but not for the 2013 Mac Pro where it's even worse) with special ROMs. And they're not able to run large volumes to soak up the cost of doing a special run.

This is probably a big reason Apple doesn't release GPUs more often, it takes them so long to make back their investment. Also why Apple has considered dropping the Mac Pro.

If I was sitting where they were, I'd actually try to get out of the GPU upgrade market entirely, and try to make it so I can just ship normal PC cards with the box and try to get back that profit margin.

Yeah, Apple could jack up their card prices to try to make money, but that just seems dumb. And it doesn't solve the problem of Apple losing money on the post-purchase upgrade market, even if they have been jacking up the card prices by $100 or $200.

Plus if selection is the problem, then it makes everything even worse. They're putting in more work on a wider range of cards that just subdivides the market more instead of growing it.

The conspiracy in these parts is that Apple makes a killing off of BTO or post purchase upgrades, and for stuff like GPUs they just don't. (RAM and CPU they likely make a killing on though.)

This is also the reason on tower Mac Pros you had third party vendors do a single Mac card, and then never be heard from again. The third party vendors all figured out the Mac GPU market was a money loser. AMD only released a Mac card every 3-5 years because they had to put in the work anyway for Apple.

If Apple wants to the Mac Pro to actually be a solid revenue stream and make customers happy, they'll be looking for an out for controlling the GPUs. Not saying they'll get there, but it's got to be on the table.

(Worth noting this is EXACTLY what they're doing for eGPU, probably because of their horrible experience on the Mac Pro.)

 

[AidenShaw]

and try to make it so I can just ship normal PC cards with the box

I hope that this message is getting to Tim and the amigos.

However, most of Apple's "innovations" are in finding creative ways to do proprietary vendor lockdown. (Try to paint the T2 chip's SSD controller in any other light.)

 

[DearthnVader]

Another Apple event, still no mMP

 

[Simon R.]

Disregard

 

[deconstruct60]

I'm 99% sure Apple has always been losing money on GPU upgrades or breaking even.

They're limited runs of reference designs (but not for the 2013 Mac Pro where it's even worse) with special ROMs. And they're not able to run large volumes to soak up the cost of doing a special run.

This seems a bit muddled. There is a difference between BTO options and "after system sale" upgrades. The Mac Pro 2013 upgrades couldn't have been worse because it wasn't even an option. Not sure how you loose money on something that your aren't selling.

Did Apple loose money because all the broken Dx00 they had to replace? Probably, but that is primarily Apple's problem, not a problem of the market.

If BTO runs is a problem then the option is drop the number of options. The MP 2013 had three options. Some evidence suggest that Apple tried to goose the "too much dilution" issue by mandating that they all be sold in pairs. If pairs was going to lead to added value in performance that would be one thing. if pairs is primarily being driving because the number of systems sold it too small to support the bundled options that is something extremely different.

But if the standard configurations are covering the sizable enough runs to get close to breakeven than that is fine. But Apple does change GPUs for the bundled/embedded options on other Macs. If the run rate for GPUs is to small how is the custom motherboard other components working out for the rest of the Mac Pro ?

The GPU upgrade stuff doesn't have to make money. In fact, for the overall market the low-mid range market is typically pretty bad. AMD and Nvidia make the bulk of their profits on their "Pro" cards. The 'race to the bottom' low-mid markets are unhealthy priced. ( the crypto miner folks have pushed all the cards up. ), but the retail ( "after system sale" ) market for GPU cards has largely sucked for years. It isn't an Apple or Mac Pro. It would suck more for Apple since the market is so narrow but management is delusional if there is some expectation that this is some big money maker. Cover the bunlded cards and be happy; anything else is just lucky lotto money (gravy).

This is probably a big reason Apple doesn't release GPUs more often, it takes them so long to make back their investment. Also why Apple has considered dropping the Mac Pro.

For the Mac Pro 'after system sale" it is going to take a long time because it ia quite likely going to get ripped by the substantive cottage industry built up around folks who just copy without paying. ROI is pretty tough we people are taking your product and not paying for it.

One of the problems was that there was only a small barrier to entry for the rouge market. If it is hardware + software solution then they aren't going to compete. If there are only 20-60K ( divide up 100K

If Apple uses a custom care to do the jump that is yet another reason to not focus on the "after system sale" upgrade market. That won't be one for several years. Apple has screwed up the Mac Pro by doing two 3-4 years runs with nothing. One of the preconditions for a upgrade market is that you have a stable base to sell into. The Mac Pro base right now is rotten and unstable to the core at this point. It is either machines so old they are on the Vintage list ( no way that is a viable market) or can't upgrade at all (MP 2013). So upgrades ? Get a grip! Win some games and then worry about the Super Bowl.


There run rate costs should be going a done a bit for GPU+Memory package boards. WIth HBMv2 a very sizable chunk of the GPU system comes prefabricated from the factory. If Apple removes the circuitry and the physical sockets for the Display output that too gets a major simplification. After that, what's left? Not alot that needs to vary much. Vega 56 and 64 are basically the same package with some parts turned off for the 56 model. Are there really huge differences between a 56 and 64 board? If the same line can make both kinds with just part differences ( no placement) the costs aren't going to be the same as doing two substantively different boards.

If I was sitting where they were, I'd actually try to get out of the GPU upgrade market entirely, and try to make it so I can just ship normal PC cards with the box and try to get back that profit margin.

Apple is also a OS vendor so they can't get out of the it completely.

Yeah, Apple could jack up their card prices to try to make money, but that just seems dumb. And it doesn't solve the problem of Apple losing money on the post-purchase upgrade market, even if they have been jacking up the card prices by $100 or $200.

Make money to cover standard configurations is different than the post-purchase market.

If the cards physically look the same Apple would always have price anchoring problems. If Apple adds value added engineering and software to the solution then charging more isn't as large a problem . As noted above that is largely how AMD and Nvidia make most of their profits (for higher than race-to-the-bottom GPU solutions). Apple can do the same thing. The Mac Pro consists of Intel CPU , 3rd party RAM , commodity storage media , etc. and Apple charges a bit more because the "integrated system that works well" effect. Same "value add" can be applied to the GPUs.

Yes. .. the moaners and groaners who want "cheaper" are going to complain but the whole Mac line up has those complainer and it is still a Fortune 500 sized business. If the bundled GPUs get to break even this post-purchase upgrade is just far more so an inventory management and sales project accuracy issue.

It isn't really jacking up prices if the run rates are smaller and the boot/drivers are solid and stable. Apple trying to slap the higher margins they do on CPU , RAM , and SSD is wrong. Or putting GPUs on task to be one equal footing as those other three. If Apple is marking 25-30% mark up on 3 out of 4 BTO options that is pretty good. At some point being too greedy is destructive.

Plus if selection is the problem, then it makes everything even worse. They're putting in more work on a wider range of cards that just subdivides the market more instead of growing it.

std PCI-e slot for the non default display GPU covers selection. eGPU will be a factor anyway so not really additional overhead for Mac Pro if cover with that system too (but with internal slot).

This is also the reason on tower Mac Pros you had third party vendors do a single Mac card, and then never be heard from again. The third party vendors all figured out the Mac GPU market was a money loser. AMD only released a Mac card every 3-5 years because they had to put in the work anyway for Apple.

As mentioned above the 3rd party folks were likely get burnt by not getting the full, expected/promised Mac Pro upgrade market because their work would get ripped off.

Again there has to be some work done to go "form over function" to use a PCI-e card. There has pragmatically evolved in the Mac GPU card market two different cards. One is fully boot/default screen enabled. The other is the "happens to work" after boot cards. They don't necessarily need the same slot. They are segmented in terms of different value adds. Apple doesn't need to 'own' the "happen to work" cards at all.

If Apple wants to the Mac Pro to actually be a solid revenue stream and make customers happy, they'll be looking for an out for controlling the GPUs. Not saying they'll get there, but it's got to be on the table.

This thing rattles down a rat hole if it turns into Apple wanting a continuing hardware profit (and revenue) stream from Mac Pro users in the out years after purchase. That's goofy. they don't do that from CPUs. why should they do it for GPUs ?

It is more critical that they continue to et Mac Pros systems sold, not sell a ton in bubble year 1 and then largely go down to nothing for 3 years and then have another bubble year. Yes some folks are going to sit on systems longer and will need to make that not too painful but not going to make up Mac Pro profits with after market GPU sales. That's the tail wagging the dog.

(Worth noting this is EXACTLY what they're doing for eGPU, probably because of their horrible experience on the Mac Pro.)

Since the PCI-e external enclosures only have standard PCI-e slots in them, there isn't much of a design choice there. They have to because that is the only option. Apple doesn't even make them. It is also part of the multi-platform ( Windows + macOS + Linux(?) ) ecosystem support for Thunderbolt. Again it isn't a sole choice for Apple design what the options are. To drag that back to the Mac Pro is pure form over function. To be a external enclusure isn't the Mac Pro's function.... it is pure form that is bring dragged back.

 

[vailr]

Again there has to be some work done to go "form over function" to use a PCI-e card. There has pragmatically evolved in the Mac GPU card market two different cards. One is fully boot/default screen enabled. The other is the "happens to work" after boot cards. They don't necessarily need the same slot. They are segmented in terms of different value adds. Apple doesn't need to 'own' the "happen to work" cards at all.

The "fully bootable" GPU card slot could also easily be re-engineered to work with a standard PC PCIe video card.
They should simply go with the industry standard UEFI "logic board" bios firmware and video card firmware.
If that were to also make the mMP a better selling machine, which I think it would, then: why doesn't Apple simply accept defeat on that point?
Or: maybe they already have, and nobody's leaked it yet?

 

[AidenShaw]

They should simply go with the industry standard UEFI "logic board" bios firmware and video card firmware

Where have I heard that before?

 

[ETN3]

Pssst...

---------------------------------------------------------------------------------------------------------------------

Mac System Software Engineer

• Job Number: 113337645
• Santa Clara Valley, California, United States
• Posted: Mar. 5, 2018
• Weekly Hours: 40.00

Job Summary
The Mac Bringup team within Apple’s CoreOS organization is seeking a talented System Software/Firmware engineer with exceptional PCIe knowledge to support future generations of Mac Systems. Weʼre looking for detail oriented engineer with exceptional problem solving skills and a passion for quality who want to define and build future Mac systems.

Key Qualifications

• Strong knowledge of PCIe technology
• Detailed working knowledge of software/firmware design and implementation.
• Expertise in C and/or C++ programming
• Previous experience with firmware and/or device driver development and debugging
• Strong Hardware/Software interaction knowledge.
• Computer architecture knowledge
• Demonstrated creative, critical thinking and troubleshooting skills
• Highly professional, with the ability to deliver solid work on tight schedules
• Outstanding written and verbal communication

Description
In this highly visible position for an individual with a strong PCIe background in software development, You will play a central role in the development of future Macs. This will include work in system software, boot firmware, and overall system architecture. Responsibilities include design, implementation, and troubleshooting of both hardware and software. To be successful in this role you must have the desire to work with hardware and software that is in a pre-production state, as well as troubleshoot issues associated with bring-up of new hardware. As a member of the Mac bringup team, you will have the opportunity to design, develop and solve challenging engineering problems in the area of PCIe technology across a broad range of Mac products.


---------------------------------------------------------------------------------------------------------------------

https://jobs.apple.com/us/search#&ss=professional&t=1&so=&lo=0*USA&pN=5&openJobId=113337645

 

[Peace]

Pssst...

---------------------------------------------------------------------------------------------------------------------

Mac System Software Engineer

• Job Number: 113337645
• Santa Clara Valley, California, United States
• Posted: Mar. 5, 2018
• Weekly Hours: 40.00

Job Summary
The Mac Bringup team within Apple’s CoreOS organization is seeking a talented System Software/Firmware engineer with exceptional PCIe knowledge to support future generations of Mac Systems. Weʼre looking for detail oriented engineer with exceptional problem solving skills and a passion for quality who want to define and build future Mac systems.

Key Qualifications

• Strong knowledge of PCIe technology
• Detailed working knowledge of software/firmware design and implementation.
• Expertise in C and/or C++ programming
• Previous experience with firmware and/or device driver development and debugging
• Strong Hardware/Software interaction knowledge.
• Computer architecture knowledge
• Demonstrated creative, critical thinking and troubleshooting skills
• Highly professional, with the ability to deliver solid work on tight schedules
• Outstanding written and verbal communication

Description
In this highly visible position for an individual with a strong PCIe background in software development, You will play a central role in the development of future Macs. This will include work in system software, boot firmware, and overall system architecture. Responsibilities include design, implementation, and troubleshooting of both hardware and software. To be successful in this role you must have the desire to work with hardware and software that is in a pre-production state, as well as troubleshoot issues associated with bring-up of new hardware. As a member of the Mac bringup team, you will have the opportunity to design, develop and solve challenging engineering problems in the area of PCIe technology across a broad range of Mac products.


---------------------------------------------------------------------------------------------------------------------

https://jobs.apple.com/us/search#&ss=professional&t=1&so=&lo=0*USA&pN=5&openJobId=113337645

This should be renamed Mac "Catch-Up" team.

Interesting about PCIe and firmware though.

 

[Mago]

Pssst...

---------------------------------------------------------------------------------------------------------------------

Mac System Software Engineer

• Job Number: 113337645
• Santa Clara Valley, California, United States
• Posted: Mar. 5, 2018
• Weekly Hours: 40.00

Job Summary
The Mac Bringup team within Apple’s CoreOS organization is seeking a talented System Software/Firmware engineer with exceptional PCIe knowledge to support future generations of Mac Systems.

---------------------------------------------------------------------------------------------------------------------

https://jobs.apple.com/us/search#&ss=professional&t=1&so=&lo=0*USA&pN=5&openJobId=113337645

Seems somebody is developing some custom PCIe System and needs special firmwre etc ...

 

[Chancha]

This should be renamed Mac "Catch-Up" team.

Interesting about PCIe and firmware though.

What I found extra interesting is that the exact term "Mac bringup/bring-up team" does not exist until 2018 Jan, according to Google.

 

[Peace]

What I found extra interesting is that the exact term "Mac bringup/bring-up team" does not exist until 2018 Jan, according to Google.

Well UEFI almost confirmed.

More from Apple's site :

Experience in UEFI firmware driver development Experience with Intel processor architecture Exposure to hardware logic analyzers, oscilloscopes Experience in macOS or iOS development Working knowledge of various scripting languages: Perl*,Python*, Bash Shell* and Make Ability to understand board schematics

 

[Aldaris]

Pssst...

---------------------------------------------------------------------------------------------------------------------

Mac System Software Engineer

• Job Number: 113337645
• Santa Clara Valley, California, United States
• Posted: Mar. 5, 2018
• Weekly Hours: 40.00

Job Summary
The Mac Bringup team within Apple’s CoreOS organization is seeking a talented System Software/Firmware engineer with exceptional PCIe knowledge to support future generations of Mac Systems. Weʼre looking for detail oriented engineer with exceptional problem solving skills and a passion for quality who want to define and build future Mac systems.

Key Qualifications

• Strong knowledge of PCIe technology
• Detailed working knowledge of software/firmware design and implementation.
• Expertise in C and/or C++ programming
• Previous experience with firmware and/or device driver development and debugging
• Strong Hardware/Software interaction knowledge.
• Computer architecture knowledge
• Demonstrated creative, critical thinking and troubleshooting skills
• Highly professional, with the ability to deliver solid work on tight schedules
• Outstanding written and verbal communication

Description
In this highly visible position for an individual with a strong PCIe background in software development, You will play a central role in the development of future Macs. This will include work in system software, boot firmware, and overall system architecture. Responsibilities include design, implementation, and troubleshooting of both hardware and software. To be successful in this role you must have the desire to work with hardware and software that is in a pre-production state, as well as troubleshoot issues associated with bring-up of new hardware. As a member of the Mac bringup team, you will have the opportunity to design, develop and solve challenging engineering problems in the area of PCIe technology across a broad range of Mac products.


---------------------------------------------------------------------------------------------------------------------

https://jobs.apple.com/us/search#&ss=professional&t=1&so=&lo=0*USA&pN=5&openJobId=113337645

*cough* shouldn’t that have been posted like before last April...

 

[deconstruct60]

Pssst...

.....
Job Summary
The Mac Bringup team within Apple’s CoreOS organization is seeking a talented System Software/Firmware engineer with exceptional PCIe knowledge to support future generations of Mac Systems. Weʼre looking for detail oriented engineer with exceptional problem solving skills and a passion for quality who want to define and build future Mac systems.
.....

PCI-e and the standard PCI-e slot physical format are two grossly different things. Every single supported Mac currently out there deployed (the vast majority with no physical PCI-e standard slot at all) use PCI-e. The relatively sloppy usage of the term 'PCI-e' to refer to a phyical form factor isn't really present in that job description at all. There is nothing particularly Mac Pro relevant there. In fact the description ends in "a broad range of Mac products". That pretty much should read all Mac products.

 

[Chancha]

*cough* shouldn’t that have been posted like before last April...

The previous engineer mistook Jony Ive's mMP prototype as the office coffee maker, filled it with water, ruining the team's entire year's worth of work.
[doublepost=1522293053][/doublepost]

PCI-e and the standard PCI-e slot physical format are two grossly different things. Every single supported Mac currently out there deployed (the vast majority with no physical PCI-e standard slot at all) use PCI-e. The relatively sloppy usage of the term 'PCI-e' to refer to a phyical form factor isn't really present in that job description at all. There is nothing particularly Mac Pro relevant there. In fact the description ends in "a broad range of Mac products". That pretty much should read all Mac products.

I was wondering the same, technically even the trackpad or the TouchBar firmware qualifies as being PCIe. The job posting's choices of words do seem to suggest the person will work across all Mac models, whether or not a Mac Pro or even any desktop Mac is central to his/her work. Though the "Mac bringup team" term is hilariously fitting to the narrative of this thread specifically.

 

[goMac]

Seems somebody is developing some custom PCIe System and needs special firmwre etc ...

There isn't much special in this listing. Every computer (Mac or PC) has custom firmware, I don't really see anything pointing to anything special or even Mac Pro here. There are a few interesting odds and ends (UEFI, "challenging engineering problems") but nothing that really screams any specific direction on the Mac Pro.
[doublepost=1522294667][/doublepost]

This seems a bit muddled. There is a difference between BTO options and "after system sale" upgrades. The Mac Pro 2013 upgrades couldn't have been worse because it wasn't even an option. Not sure how you loose money on something that your aren't selling.

I think that's kind of the point. Apple can't lose money on something they aren't selling.

If the run rate for GPUs is to small how is the custom motherboard other components working out for the rest of the Mac Pro ?

That's my point. Any bean counter looking at the Mac Pro right now is going to be screaming "custom components bad!"

I mean, the motherboard is kind of baked into the cake, the Mac Pro is going to have a custom board just like other Macs. But that's also why Apple very rarely touched the motherboard on the cMP. Yeah, they could have added a SATA3 controller but what a pain in the ass to spend more money touching that board.

The idea that Apple makes money off their proprietary GPUs is backwards. They'd actually make more money off of an off the shelf GPU, especially if they kept charging $100 extra for the privilege of having them install it (especially with the crypto rush on retail GPUs.)

Again, Thunderbolt might hold them back. But for profitability, a off the shelf GPU with off the shelf firmware is the way to go.

The GPU upgrade stuff doesn't have to make money.

It doesn't. But it's also the reason Apple doesn't do GPU upgrades.

If Apple uses a custom care to do the jump that is yet another reason to not focus on the "after system sale" upgrade market. That won't be one for several years.

It's a catch-22. Apple, for all the reasons you've highlighted, does not want to be involved in the aftermarket upgrade market. You're absolutely right. But they also know the Mac Pro won't be successful without users being able to do aftermarket upgrades.

The answer? Build a Mac Pro that takes industry standard, generic PC upgrades so they can both sell a system that is upgradable, but stay out of the aftermarket upgrade market as well. They get what they want, users get what they want, everyone wins.

Again, there are technical reasons why that may not happen, but someone at Apple has got to be thinking that. Especially in since they're all sitting around asking why Mac Pro users are so happy, and why they've been unable to ship timely upgrades.

There run rate costs should be going a done a bit for GPU+Memory package boards. WIth HBMv2 a very sizable chunk of the GPU system comes prefabricated from the factory. If Apple removes the circuitry and the physical sockets for the Display output that too gets a major simplification. After that, what's left? Not alot that needs to vary much. Vega 56 and 64 are basically the same package with some parts turned off for the 56 model. Are there really huge differences between a 56 and 64 board? If the same line can make both kinds with just part differences ( no placement) the costs aren't going to be the same as doing two substantively different boards.

The costs for Apple doing a custom GPU:
- Apple has to do a special, limited run of the GPU. With the 2012 Mac Pro and earlier that wasn't awful, they were really just doing a slightly tweaked reference card. With the 2013 this has become much worse because the design is very custom.
- They have to write the custom firmware. I don't know if they're doing it or they're passing it off to AMD, but the custom ROM is probably $200k at least in engineer and QA time. Even though the ROM is cheap to flash, it's going to take time and money to write and test, and Silicon Valley engineers aren't cheap.
- Then the driver has to be written. I don't know if they're paying AMD to write it (the driver is definitely being written at AMD) but you're probably looking at another $200k.

It isn't really jacking up prices if the run rates are smaller and the boot/drivers are solid and stable. Apple trying to slap the higher margins they do on CPU , RAM , and SSD is wrong. Or putting GPUs on task to be one equal footing as those other three. If Apple is marking 25-30% mark up on 3 out of 4 BTO options that is pretty good. At some point being too greedy is destructive.

Apple could have it both ways. They could ship a standard PC card AND mark it up especially with no one able to buy at retail for reasonable prices for the foreseeable future.

But if the Mac Pro doesn't have upgradable GPUs we're all wasting our time, and I think Apple knows that. Changing the markup doesn't solve that problem, especially when the markup is really just for absorbing their extra engineering costs.

Again there has to be some work done to go "form over function" to use a PCI-e card. There has pragmatically evolved in the Mac GPU card market two different cards. One is fully boot/default screen enabled. The other is the "happens to work" after boot cards. They don't necessarily need the same slot. They are segmented in terms of different value adds. Apple doesn't need to 'own' the "happen to work" cards at all.

Two kind of slots is an option. It doesn't solve everything for Apple but it would be workable. You can have an onboard card, and especially if you're not using a Thunderbolt display, you could just plug in to your PCIe card.

I think there is a good chance the boot screen problem goes away completely because Apple will adopt standard UEFI. Boot screens are a completely solvable problem on Apple's end.

(I'm pretty sure new Macs have already adopted UEFI GOP so boot screens may already be a solved problem. I think someone mentioned that there is still a bit of custom functionality, but that's solvable.)

This thing rattles down a rat hole if it turns into Apple wanting a continuing hardware profit (and revenue) stream from Mac Pro users in the out years after purchase. That's goofy. they don't do that from CPUs. why should they do it for GPUs ?

They wouldn't. That's the point. They need an upgradable machine but they don't want to be the gatekeeper for upgrades.

The cMP, being an early EFI machine, forced Apple to be the GPU gatekeeper because they were the only ones who made EFI GPUs. Same with the Power Mac when it was the only consumer box with Open Firmware.

Now EFI GPUs are commercially available. They're not forced to do GPUs anymore.

Like we've talked about, the only real wildcard is Thunderbolt and how that interacts with the GPU. But on the firmware side, Apple isn't forced to do anything custom any more.

Since the PCI-e external enclosures only have standard PCI-e slots in them, there isn't much of a design choice there. They have to because that is the only option. Apple doesn't even make them. It is also part of the multi-platform ( Windows + macOS + Linux(?) ) ecosystem support for Thunderbolt.

Oh sure, but if Apple wanted to lock that market up, they'd make eGPU only work with the "Apple eGPU" that they sell for an absurd markup, like MacVidCards was paranoid about back in the day. They don't have to support the standards.

But again, as I've been saying, they don't want to because they don't want to be a gatekeeper for GPU upgrades. They have no interest. It doesn't make business sense for them to control eGPU upgrades just like it doesn't make sense for them to control GPU upgrades on the Mac Pro. I doubt we'll ever even see an Apple branded eGPU, especially with them out of consumer displays. Maybe the new pro display will be some crazy design with a built in eGPU but I kind of doubt it.

I think a lot of this comes down to if you think the Mac Pro is basically going to be the iMac Pro without the built in display. A locked down, not upgradable box with the only extra feature being you can plug in your own displays. If that's what the Mac Pro is I think Apple is wasting it's time. I think they know that too. At least I hope they do. But if that's all the Mac Pro is, it also doesn't make sense we'd be waiting this long.

If that's all the Mac Pro is I might even end up making my way over to the iMac Pro forums, cause really, what's the point. If I can't upgrade my GPU I might as well just get an iMac Pro. Or just build a PC.

 

[deconstruct60]

The "fully bootable" GPU card slot could also easily be re-engineered to work with a standard PC PCIe video card.

If you move the goal posts and ignore signification constraints then sure. There are three factors. Booting in Apple's secure boot environment, integration with Thunderbolt, and suitability to the system being built.

There are Rube Goldberg solutions for 1 TB socket workstations that Dell/HP/etc use but they get unwieldy when scale up to 4 ( matching the iMac Pro or MBP 15"). A deliberate socket that provisioned 4-6 video streams in a single insert connection aligned with the single insert PCI-e connection would be cleaner and less error prone. Four different short cables that hooked four different sockets would be quite conrived even above the 1 socket solutions.

Why would a Mac Pro be incapable of doing computation and display things that most Macs can do? Shouldn't the Mac Pro do more than the other ones? That used to be a differentiator.

re-enginering is fundamentally different from picking your solution beforehand and then mutating all of the constraints around to fit that preselected solution. re-egineering is fitting new solution the context that you have, not changed the context to fit what is already exists.

They should simply go with the industry standard UEFI "logic board" bios firmware and video card firmware.

BIOS is probably one of the reasons it isn't going to happen. Apple probably really didn't want to screw around with BIOS. In 2005-6 EFI came out and Apple left OpenFirmware for EFI. UEFI didn't arrive until 2007 (after Apple had already started to evolve their EFI implementation. ). All along though BIOS has been dragged along with BIOS compatibility more. 10 years later this crutch is still around. Intel is only one announced getting off in 2020

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

That is almost 15 years after EFI. The excuse has been all the deployed systems that use BIOS crutches to get by in boot and operational modes. Well Apple has the same stuff. There are all these older Apple EFI systems that have dependences on older implementations of EFI. As long as the mainstream market keeps kicking the can down the road Apple will probably keep kicking the can down the road on fully transitioning to UEFI. They had added bits and pieces where not too much work but little motivation to take on the pain of a full transition when the mainstream market doesn't want to do it either.

Secure booting was somewhat later glued into UEFI and I suspect it too may need some updates before Apple would find it worthwhile to completely merge.

If that were to also make the mMP a better selling machine, which I think it would, then: why doesn't Apple simply accept defeat on that point?
Or: maybe they already have, and nobody's leaked it yet?

If Apple were to spend significant R&D to completely maximize lower the boundaries to Hackintoshes that would increase the number of Macs sold. I doubt anyone at the mid-upper levels of Apple executives will buy into that notion. It is a nice fantasy some customers , but it won't.

 

[Peace]

Remember when devs showed a graphic with fast os switching circa wwdc 2007 ?

good times.