MP All Models Waiting for Mac Pro 8,1

[deconstruct60]

God knows when we will see a Ice Lake based W chip, that could slip into 2021.

Intel can't really afford to wait until 2021. The high end Ryzen and the upcoming Threadripper will eat them alive in the workstation market if they snore through all of 2020 ( barring AMD going back to shooting themselves in the foot. AMD has a 'tick' (mostly shrink) lined up for 2020. By 2021 AMD will be on their next 'tock' . )

What we probably won't see is a core count increase. May even drop 4-6 cores for Ice Lake W (e.g., max 14 core count) . ( if uptick 18% per core then 10 cores is a 180% uptick .. almost two. ) . Couple that with some less arrogant pricing and they'll "survive' 2020 to get to 2021 with something that is "good enough".

Intel's immediate customers aren't happy with Intel's lack of execution. Punting products into the future each year as a excuse is suppose to be what Intel is turning around.

Even with all of the 10nm FUBAR they still managed to get updates out the door with the "W class" die. In 2018 it was just tweaked i9 caches and clock ( and only a hint at the 3647 switch ), but they did something. It was no Rip Van Winkle act.

Even if 10nm++ means have to drop core count they'll just roll with that because it shows progress. And that they can better protect the 8-14 core + high end bandwidth workstation space in the interim.

And that would probably be OK for the iMac Pro on Apple's scorecard.

 

[bpeeps]

Joking aside, I think the 8,1 will come out as a 2020 model with Cooper Lake Xeon W and will be visually identical to the 7,1. That will be PCIe 4.0.

Think very much like the 4,1 to 5,1 models.

You think apple is going to release a 2020 Mac Pro after all the money, r&d, and upgradeability they invested in the. 2019 Mac Pro? That’s a ridiculous notion.

 

[deconstruct60]

Think very much like the 4,1 to 5,1

You think apple is going to release a 2020 Mac Pro after all the money, r&d, and upgradeability they invested in the. 2019 Mac Pro? That’s a ridiculous notion.

I think yours not getting the point that the 4,1 and 5,1 share pragmatically the exact same motherboard , case and basic infrastructure . A rogue firmware upgrade of a 4,1 turns it into essentially a 5,1.

There was no huge R&D cost for Apple from the 2009 [4,1] to 2010 ( and 2012 ) [5,1] models . Apple costed for 4 years on that baseline R&D investment .

If Intel hands Apple a W-33xx series processors the fit in the same sockets that the W-32xx fit into it is a relatively super cheap upgrade . Apple would change the Mac Pro model number just like they did in 2010.


A whole new motherboard and whole new set of MPX modules? Highly likely not ( Apple recently taking 4 years to fix a keyboard and two years of ‘dog ate my homework’ delays on new Mac Pro ... no way unless they farmed the whole thing out to an external group. Has nothing to do with money or ability to invest at all and everything to do with Apple being able to walk and chew gum at same time in Mac space . ) . However, that is exactly not like the 2009 to 2010 transition at all .

 

[Slash-2CPU]

You think apple is going to release a 2020 Mac Pro after all the money, r&d, and upgradeability they invested in the. 2019 Mac Pro? That’s a ridiculous notion.

In all likelihood, moving to Cooper Lake will require little, possibly no, physical redesign of existing motherboards. In that case, it would be more ridiculous to keep selling a Cascade Lake system when everyone else just stepped across to Cooper Lake for almost no R&D cost and a clock speed bump and/or core count increase for the same price.

If it's a move to a new socket and a change in voltage requirements for a single-digit percentage performance increase, then that won't happen.

 

[bpeeps]

In all likelihood, moving to Cooper Lake will require little, possibly no, physical redesign of existing motherboards. In that case, it would be more ridiculous to keep selling a Cascade Lake system when everyone else just stepped across to Cooper Lake for almost no R&D cost and a clock speed bump and/or core count increase for the same price.

If it's a move to a new socket and a change in voltage requirements for a single-digit percentage performance increase, then that won't happen.

That’s not the point, it’s gonna be a while before Apple recoups the cost they put into the 2019 Mac Pro. The cost of a 2020 Mac Pro is irrelevant. They could have just as easily stuffed some new internals in the iMac in 2018 and released a new configuration, but they didn’t. iMacs have been on a 2 year cycle since 2015 when they could have been easily updated every year for very little cost.

 

[deconstruct60]

In all likelihood, moving to Cooper Lake will require little, possibly no, physical redesign of existing motherboards. In that case, it would be more ridiculous to keep selling a Cascade Lake system when everyone else just stepped across to Cooper Lake for almost no R&D cost and a clock speed bump and/or core count increase for the same price.

Cooper Lake is highly likely not going to bring any core count. 14nm++ basically the same as Cascade lake. ( maybe Intel adds another '+' on the end but basically same stuff different day. )

If go back to the chart in post #23 you'll that the 4-8 socket model caps out at 26 cores. That is actually two less than Cascade lake. The other Cooper Lake variant is getting to 48 cores, which extremely likely by throwing two 24's into a even bigger and hotter package than before.

The 4-8 socket model also have more UPI links. That is partially so the dual package link ups have something more to use internally.

Since it is just 14nm++ ( or +++++ ), it isn't likely going to clock faster at all at the same TDP. Intel could run the whole product line hotter, but that probably wouldn't 'sit' very well with the Mac Pro. ( which is running its GPUs even hotter also. ). Probably talking maybe a couple 100 MHz bump.

Fo the 4S-8Socket crowd it is the last gasp so they'll take it. For the 2S crowd it is far more so a path the Ice-Lake than some huge expected performance jump (for general purpose workloads )

If it's a move to a new socket and a change in voltage requirements for a single-digit percentage performance increase, then that won't happen.

That is about all that Cooper Lake is really going to do. Unless running Google (AI ) targeted BFloat opcodes , better Optane DIMMs , and/or extremely high memory I/O rates. there isn't much changed over SP 2nd Generation (Cascade). Neither one those two are in historical Mac Pro wheelhouse. [ Apple probably won't touch Optane DIMMs so those are out also. And Intel isn't likely to bring it down to W-series also. ]

If Intel puts Cooper Lake into 3647 socket for a W-33xx variant, then the extremely high memory I/O goes out the window. So it would be bfloat and some <5% bump. But if simply just a cheap firmware upgrade Apple probably will take it and bump the model number.

Huge cloud service vendors that have substantive AI inference workloads will be happy. Same with those using Optane DIMMs. There are lots of folks who will buy these. It just isn't primarily targeted at Apple's narrow area.

 

[Nugget]

too early but...

The New Mac Pro Is Almost Certainly Coming!

 

[Slash-2CPU]

That’s not the point, it’s gonna be a while before Apple recoups the cost they put into the 2019 Mac Pro. The cost of a 2020 Mac Pro is irrelevant. They could have just as easily stuffed some new internals in the iMac in 2018 and released a new configuration, but they didn’t. iMacs have been on a 2 year cycle since 2015 when they could have been easily updated every year for very little cost.

Your logic is almost a corollary of the sunk cost fallacy. The focus in business is not ever recouping but pursuing further gains. Recouping as a logical goal makes sense for small businesses or households, but even then it's purely relative and somewhat arbitrary.

Comparing life cycle of a heavily-integrated all-in-one appliance to that of what is now priced as near enterprise-grade hardware isn't exactly apples to apples.

 

[skippermonkey]

Jeez, I haven't even specced my new Mac Pro yet, let alone bought it or used the damn thing... and you're all wittering on about whatever the phrase is for pre-vapourware!

 

[Derived]

Uhm... yeah. That'd be nice, but taking into account the Apple of the past decade, the next update to Mac Pro will be around 2025 or never. "Okay, here's your Cheesegrater, now it looks more like one. There ya go! Now please shut up and go away for a few years. Once the hysteria reaches another crescendo we'll release another Mac Pro."

If you're hesitant about buying nCMP then wait a few months for firmware updates to everything that's critically wrong (T2 chip is primary concern for me, and being able to boot around it into PCIe RAID0). After that, I strongly suspect there will be a bunch of cards tested/released which will all work with nCMP, and ... absolute silence/nothing, Apple goes back to sleep and forgets about Mac Pro just like they always do. I somehow do not see a renascence of Mac Pro ... I see a small handful of people and a bunch of studios buying them, and if there's enough of those people who buy it, they may update it after 3 years since it's back to being a big tower where they can re-use the same design.

In my humble opinion, thinking there will be an updated Mac Pro in 2020 or 2021 is beyond wishful thinking.

Why is this your opinion? They've completely rebooted their pro product team, and are making a large push to start regaining their ground in this space. Your "humble opinion" surely must be based on something? It's already been re-hashed 10,0000x why they couldn't really do anything with the cylinder MP...they thought more performance/watt, and lower power usage in general, were coming from chip manufacturers...that didn't happen. They thought 500 watts and just barely enough cooling for 500 watts would be enough...it wasn't. And the silicon needed to iterate that design never materialized. They screwed up. End of story. They didn't just decide to spite the customer base, that would be the most insane course of action any company has ever taken. There's less than zero upside in them doing so, as we can clearly see by all the anger and frustration on this board over the last few years. Can we finally move past the "Apple just really hates the pros and secretly designed a bad product to spite them, piss them off and force them to leave" conspiracy and just, you know, accept they effed up and appreciate the effort they're making now? Seems far more productive to me. The new machine lends itself to easy, iterative upgrades. I see no reason at all why this wouldn't happen, especially after what was likely a several hundred million dollar effort.

 

[bpeeps]

Your logic is almost a corollary of the sunk cost fallacy. The focus in business is not ever recouping but pursuing further gains. Recouping as a logical goal makes sense for small businesses or households, but even then it's purely relative and somewhat arbitrary.

Comparing life cycle of a heavily-integrated all-in-one appliance to that of what is now priced as near enterprise-grade hardware isn't exactly apples to apples.

And you think releasing a Mac Pro in 2020 chases further gains. I have to laugh.

 

[AidenShaw]

And you think releasing a Mac Pro in 2020 chases further gains. I have to laugh.

Part of re-establishing a commitment by Apple for pro hardware would be to keep things up to date. If there are new components that are faster - nothing will demonstrate commitment better than prompt upgrades.

 

[jinnyman]

Well. 7.1 Mac Pro is nice and good, but I want it to fail financially so that Apple may come up with a new Mac pro with an appropriate price tag for people like me. I really wanted to get a real desktop mac and was willing to pay upto $4,000, but alas.

Good luck to all of you who would get this generation of Mac Pro! Until then, I will enjoy my hackintosh!

 

[CWallace]

Well. 7.1 Mac Pro is nice and good, but I want it to fail financially so that Apple may come up with a new Mac pro with an appropriate price tag for people like me. I really wanted to get a real desktop mac and was willing to pay upto $4,000, but alas.

If it fails financially Apple will just kill the model and the iMac Pro will anchor the top of the line (which I continue to believe was the original plan before they were publicly shamed into producing the 2019 Mac Pro).

 

[ZombiePhysicist]

Why is this your opinion? They've completely rebooted their pro product team, and are making a large push to start regaining their ground in this space. Your "humble opinion" surely must be based on something? It's already been re-hashed 10,0000x why they couldn't really do anything with the cylinder MP...they thought more performance/watt, and lower power usage in general, were coming from chip manufacturers...that didn't happen. They thought 500 watts and just barely enough cooling for 500 watts would be enough...it wasn't. And the silicon needed to iterate that design never materialized. They screwed up. End of story. They didn't just decide to spite the customer base, that would be the most insane course of action any company has ever taken. There's less than zero upside in them doing so, as we can clearly see by all the anger and frustration on this board over the last few years. Can we finally move past the "Apple just really hates the pros and secretly designed a bad product to spite them, piss them off and force them to leave" conspiracy and just, you know, accept they effed up and appreciate the effort they're making now? Seems far more productive to me. The new machine lends itself to easy, iterative upgrades. I see no reason at all why this wouldn't happen, especially after what was likely a several hundred million dollar effort.

Change out the "pro market" in your above assertion and replace it with "enterprise market" and see why many apple fans/users have the "once bitten, twice shy" feeling about everything apple works on that isn't genuinely a priority at apple.

Heck, right now the only mac that is even semi-regularly updated is the MacBook Pro (which is not a pro machine, no machine without user upgradable storage is truly a 'pro' machine) and all the other macs have been updated ad-hoc at best. Maybe if Apple shows a genuine trend and commitment (say more than 5 years) at any of this stuff, perceptions might start to change. But for right now, all they seem to care about is the iPhone (when you judge them by actions/products shipped and not PR/marketing/words/fluff).

In these forums, the feelings of upgrading to the 7,1 vs an 8,1/9,1 are about 50/50 at the moment...

https://forums.macrumors.com/threads/poll-when-to-buy-a-revised-cheese-grater-mac-pro.2189615/

Not sure how this forum maps out to the wider market though.

 

[jinnyman]

If it fails financially Apple will just kill the model and the iMac Pro will anchor the top of the line (which I continue to believe was the original plan before they were publicly shamed into producing the 2019 Mac Pro).

If it fails, then it kinda proves that 7.1 Mac Pro made for an entreprenuer or big encoding company is not really meant for financial success. Apple deliberately has given up on enthusiasts and since there's no product for us, why would the end of Mac Pro worry me anyway? Well let's see how well 7.1 Mac Pro does. I'm probably going to laugh when the price for upgrade is announced. They are going to be all

 

[CWallace]

Change out the "pro market" in your above assertion and replace it with "enterprise market" and see why many apple fans/users have the "once bitten, twice shy" feeling about everything apple works on that isn't genuinely a priority at apple.

This is not at all meant as a disparaging response to your comment, but the "enterprise market" (as in Dell, HP and Lenovo) has been what many on this forum consistently state Apple should aspire to be much more like.

If it fails, then it kinda proves that 7.1 Mac Pro made for an entreprenuer or big encoding company is not really meant for financial success.

While workstation shipments are said to have seen strong growth in 2018 compared to 2017, they still account for around 5 million units (however, this includes 2 million units priced under $1000 so that would be models using components like Core family CPUs and lower-end GPUs). HP, Dell and Lenovo still see decent revenues on those volumes so Apple (which will likely ship mostly $10,000 or more systems) might still turn a basic profit even if they only ship tens/scores of thousands of units a year.

Apple deliberately has given up on enthusiasts and since there's no product for us, why would the end of Mac Pro worry me anyway? Well let's see how well 7.1 Mac Pro does. I'm probably going to laugh when the price for upgrade is announced.

I'll be interested to see what Apple charges compared to what Dell and HP charge for similar parts. Those OEMs are not shy about their markups on CPUs, memory and storage.

 

[ZombiePhysicist]

This is not at all meant as a disparaging response to your comment, but the "enterprise market" (as in Dell, HP and Lenovo) has been what many on this forum consistently state Apple should aspire to be much more like.

Very much agree with your comment above. But perhaps it's extra damning in that I was referring to Apple's failures in enterprise. That Apple has made and broken so many promises to serve the enterprise market, that no one would dare take them seriously. Their xserve, xraid, mac os server product abandonments are just the tip of the iceberg of Apple enterprise failures.

 

[Derived]

I don't think it's accurate to say they "broke promises" - unless there's specific things I'm unaware of. No company has an obligation to sell you a product forever. After a short while, no one bought the Xserve, or OS X Server, so that's why they got axed. They're not going to keep spending money on products no one wants. And, frankly, when everyone in "enterprise" refuses to turn away from the Microsoft yellow brick road, largely, why is it Apple's fault? They tried to make great products in that space, everyone said "Thanks, but no thanks", and that was that. What else were they to do?

In terms of the Mac Pro specifically, I think, again, it's quite clear why they didn't update it for such a long time between the cylinder and the new design. I think it should have taken less time, but it's not at all as if they just sort of "forgot" about it, or as if they just didn't care. That's obviously not the case, and we know that. They just took a long time to develop the new machine. That's a critique, but it's not at all the same thing as Apple just deciding not to make one ever again...which for some very, very strange reason that I cannot possibly comprehend, a great deal of people on this site seem to think is still likely to happen.

 

[ZombiePhysicist]

I don't think it's accurate to say they "broke promises" - unless there's specific things I'm unaware of. No company has an obligation to sell you a product forever. After a short while, no one bought the Xserve, or OS X Server, so that's why they got axed. They're not going to keep spending money on products no one wants. And, frankly, when everyone in "enterprise" refuses to turn away from the Microsoft yellow brick road, largely, why is it Apple's fault? They tried to make great products in that space, everyone said "Thanks, but no thanks", and that was that. What else were they to do?

In terms of the Mac Pro specifically, I think, again, it's quite clear why they didn't update it for such a long time between the cylinder and the new design. I think it should have taken less time, but it's not at all as if they just sort of "forgot" about it, or as if they just didn't care. That's obviously not the case, and we know that. They just took a long time to develop the new machine. That's a critique, but it's not at all the same thing as Apple just deciding not to make one ever again...which for some very, very strange reason that I cannot possibly comprehend, a great deal of people on this site seem to think is still likely to happen.

Not every broken promise has to be an explicit one like promising to open source iMessages/facetime and failing to do so. They totally broke many promises to enterprise. Gutting key services in Mac OS X Server for those that RELY on a server is a broken promise IMO. Having business spend substantial amounts on Xserve and XRaid products to just have the rug yanked out under them caused many an IT professional to at best look stupid, and many lost jobs over trusting that Apple would make a commitment that smaller business could rely on. Same with OSX server. And there is a long history of this going back to the 80s where apple would start some enterprise initiative only to grow bored and cancel it a couple of years later.

Apple's reputation in enterprise is garbage, and it's actually well earned IMO. As always, YMMV.

 

[AidenShaw]

I'll be interested to see what Apple charges compared to what Dell and HP charge for similar parts. Those OEMs are not shy about their markups on CPUs, memory and storage.

Let's be sure to compare *selling* price, not *list* price.

Most HP/Dell systems are sold with substantial discounts from list, even for single retail sales. Corporate accounts get even lower prices.

 

[Blair Paulsen]

FWIW, I don't see a major update to the 7,1 for a few years. They likely will add a W-33xx variant to the BTO list, but why abandon the current socket/mobo design until the chip fabs nail 10 and 7nm production?
I can understand why regular updates are needed for sealed systems, but a slotbox? I'd rather Apple wait until the juice is worth the squeeze.
IMO, once the process shrink is farther along, design options for workstations will be worth revisiting. Until then...

 

[Ursadorable]

No Wi-Fi 6 on Cascade Lake SP

Wait, wut? Who uses Wi-Fi on high end professional grade hardware?

As an aside, I wonder how much slower it's going to run with all the necessary patches for Intel CPU hardware bugs verses them using a Threadripper.

 

[Krevnik]

FWIW, I don't see a major update to the 7,1 for a few years. They likely will add a W-33xx variant to the BTO list, but why abandon the current socket/mobo design until the chip fabs nail 10 and 7nm production?
I can understand why regular updates are needed for sealed systems, but a slotbox? I'd rather Apple wait until the juice is worth the squeeze.
IMO, once the process shrink is farther along, design options for workstations will be worth revisiting. Until then...

The 2003 PowerMac G5 chassis survived a full decade, even through the Intel switch. I see the same thing with the 2019. If anything, Apple is best served by amortizing costs around the MPX/Thunderbolt work across multiple refreshes and drawing folks into a regular update cycle if they can. Making folks wait isn't going to help long-term sales, and makes recouping costs harder, not easier.

I agree that they shouldn't touch the mobo until a new socket and matching chipset comes out, but Ice Lake Xeons are "Soon(tm)", and those will bring a new socket and chipset. It really depends if Ice Lake lands quickly enough that it's worth just leap-frogging to it instead of refreshing with Cooper Lake, IMO. If that happens, I expect the mobo layout to be identical, but with the new socket and chipset. And that's what should happen, IMO.

Apple would be well served to try to standardize on the slot layout they've picked with the 7,1 and treating it like an ATX mobo: stuff generally needs to be in the same place, but it still needs to be revisited when new chipsets land so it can support them. I suspect that may have been the intention with how it is laid out. It's about the most flexible/open layout I've ever seen come out of Apple. My old PowerMac 8600 weeps at how straight-forward the 2019 is laid out.

Part of re-establishing a commitment by Apple for pro hardware would be to keep things up to date. If there are new components that are faster - nothing will demonstrate commitment better than prompt upgrades.

Yes. Especially with a "slotbox" being able to be refreshed based on what's available. Cooper Lake and Navi are clear upgrades that they can do in the next 12 months to show that commitment. They just need to get out of the mindset that a refresh needs to be whiz-bang above and beyond (something Laptops and the iMac do need to be at times), and they could be fine here. They just need to make it so that folks aren't having to come in the middle of a larger cycle and feel like they need to wait.

 

[Blair Paulsen]

Another potential plus for longevity, is the large socket size out of the gate. If, as most of us suspect, process shrink issues are worked out over time - the next gen chips should have enough space to fit. Yes, there are other issues besides space - PCIe3 for example - but at least the option is likely to be there...