Actually what Apple did was remove the base 4-core D300 system.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 7,1 Waiting for Mac Pro 7,1 A1991 (no more)
- Thread starter Mago
- Start date
-
- Tags
- macpro 2016
- Sort by reaction score
Actually what Apple did was remove the base 4-core D300 system.
And drop the price of what was above it by $1000. So the new base line system offers more performance for the same price. And they lowered the price of CPU and GPU upgrades (as well as lowering the price of SSD upgrades in October).
It is still not a great value, but it's a better value than it was yesterday.
I'd generally put money on no SATA either, although having one or two bays would be a concession to people with larger storage needs. Whatever the new Mac Pro looks like I doubt it's going to satisfy the people who want to stick 16TB of storage in their computer. But pricey (and very fast) PCIe storage will continue to get more affordable and with larger capacities. If there's room to swap on the new version I don't imagine it will create that much of a hue and cry.
I'd aim fall for Mac minis and possibly WWDC for the iMacs, personally.
At this point they might as well get the newest MBP internals into a Mini revision.
With the move to APFS, I expect Apple's longer-term plan is to scrap spinning metal in the iMac.
Depending on what Apple has planned for the iMac design, we will probably still see Fusion in the Late 2017 upgrade (but now as the entry-level as opposed to just HDD). Pure SSD locally is still too expensive so Fusion (though hopefully only 2TB and 3TB models with the 128GB SSD) would still be a way to provide large storage for home users while keeping costs lower and leveraging APFS.
Depending on what Apple has planned for the iMac design, we will probably still see Fusion in the Late 2017 upgrade (but now as the entry-level as opposed to just HDD). Pure SSD locally is still too expensive so Fusion (though hopefully only 2TB and 3TB models with the 128GB SSD) would still be a way to provide large storage for home users while keeping costs lower and leveraging APFS.
HDDs are pretty much only good for backup drives at this point in time. I keep important stuff on the boot disk of both my laptop and my cMP, and everything else gets relegated to the multitude of spinny drives I have. Kudos if they make the switch to all SSD drives, something that they should've done years ago.
I get the definite impression Apple only just realized how badly they screwed up, like maybe just in the last month or so, and have barely even begun work on a redesign. It's crazy, since at least two years ago they should've figured out the nMP was a failure and went back to the drawing board, but that's the impression I get. I'm still not convinced goMac has inside info, but I agree with him that this doesn't seem like a stall strategy. This whole damage control PR event has the scent of panic about it. Apple admitting fault is extremely rare, and to do so with a "round table discussion" with five of their biggest fans in the tech press, and everything else about this, just seems like an effort to get some kind of closure after a "fustercluck" as goMac so eloquently put it. Apple's trying to move forward as best they can after things have gone horribly wrong behind the scenes.
So, I think it's way too early to start guessing what Apple's going to make next, and I'd read quotes from Schiller et al. with a hefty grain of salt. I can't help but wonder if this announcement wasn't more about signalling their engineers with a public statement of principles than an effort to stem the bleeding as pros look elsewhere for machines they need.
It might be very interesting to see if anyone makes a public statement at WWDC about what they're intending to do, even if they have nothing to show us yet. WWDC 2017 is in early June. That gives them two more months to nail down some kind of intended specs. They're not going to give us any, of course, but they might be able to hint at least about their intentions for internal storage, PCIe slots, 1 vs. 2 CPUs, etc. If they talk about flexibility in configuration, and upgrades to things like GPUs, I would take those as good signs.
I, meanwhile, am quite happy with my Hackintosh. I still have my 5,1 going as well for legacy software. I'm still full speed ahead on switching my workflow to Linux as an alternative if things continue to go poorly with the Mac. If Apple has a good 7,1 next year, and a good 8,1 a year or two after that, maybe I'll be willing to come back to Apple. I'll need to be impressed with Apple's care and attention, however, and confident Apple "gets it" again.
So, I think it's way too early to start guessing what Apple's going to make next, and I'd read quotes from Schiller et al. with a hefty grain of salt. I can't help but wonder if this announcement wasn't more about signalling their engineers with a public statement of principles than an effort to stem the bleeding as pros look elsewhere for machines they need.
It might be very interesting to see if anyone makes a public statement at WWDC about what they're intending to do, even if they have nothing to show us yet. WWDC 2017 is in early June. That gives them two more months to nail down some kind of intended specs. They're not going to give us any, of course, but they might be able to hint at least about their intentions for internal storage, PCIe slots, 1 vs. 2 CPUs, etc. If they talk about flexibility in configuration, and upgrades to things like GPUs, I would take those as good signs.
I, meanwhile, am quite happy with my Hackintosh. I still have my 5,1 going as well for legacy software. I'm still full speed ahead on switching my workflow to Linux as an alternative if things continue to go poorly with the Mac. If Apple has a good 7,1 next year, and a good 8,1 a year or two after that, maybe I'll be willing to come back to Apple. I'll need to be impressed with Apple's care and attention, however, and confident Apple "gets it" again.
Godot finally made an appearance! Unfortunately he is empty handed.
What I was able to infer from the very carefully crafted Apple speak:
1) Apple at least made some attempt to update the MP6,1, but could not make it work.
2) Apple only relatively recently made the commitment to hit the MP reset button.
3) Apple is again trying to over engineer a super fantastic brain melting solution to something that just requires a simple but reasonably elegant approach, that can deliver the power and configuration choices appropriate for each "pro" users need.
4) Apple must have practiced talking with their tongues firmly embedded in their cheeks when they mentioned frequent update capability.
Item 4 seems the most disingenuous claim due to Apple being all in on Thunderbolt. As long as they insist on all GPU output via Thunderbolt, that is going to mean custom Apple GPUs. And that means limited choices, and infrequent updates. Apple is dangling a thread of hope in front of the faithful, but as long as the MP has existed frequent GPU updates have never been the case; even with the cMP. Even with cMP Apple supported GPUs were a rarity, and updates were glacial.
Banking on that changing for the MP7,1, just seems foolhardy at best.
What I was able to infer from the very carefully crafted Apple speak:
1) Apple at least made some attempt to update the MP6,1, but could not make it work.
2) Apple only relatively recently made the commitment to hit the MP reset button.
3) Apple is again trying to over engineer a super fantastic brain melting solution to something that just requires a simple but reasonably elegant approach, that can deliver the power and configuration choices appropriate for each "pro" users need.
4) Apple must have practiced talking with their tongues firmly embedded in their cheeks when they mentioned frequent update capability.
Item 4 seems the most disingenuous claim due to Apple being all in on Thunderbolt. As long as they insist on all GPU output via Thunderbolt, that is going to mean custom Apple GPUs. And that means limited choices, and infrequent updates. Apple is dangling a thread of hope in front of the faithful, but as long as the MP has existed frequent GPU updates have never been the case; even with the cMP. Even with cMP Apple supported GPUs were a rarity, and updates were glacial.
Banking on that changing for the MP7,1, just seems foolhardy at best.
For "power users" and "professionals", I agree. But I expect a fair bit of iMac sales go to households and education where they want everything inside or are not comfortable with having external secondary storage they need to manage.
Eventually iMacs will go all-SSD just as the laptop lines have as prices come down (while allowing Apple to maintain their high margin percentages). But for the near-term, I still think Fusion will have a role to play for the entry-level iMac models.
A 1TB SSD isn't that expensive, maybe $100 for a Samsung model, mark it up as a $200 option (in typical apple fashion) and that's more than their average profit margin, without breaking the bank for the average iMac buyer. Most people don't go beyond 1TB anyway.
Entry level stuff should be relegated to the Mac mini in my opinion. Fusion drives have a market there where you're looking at a sub-$500 computer and every penny counts.
This is largely false and would be a mistake for Apple.
If the only thing they've learned from the nMP debacle is that instead of 1 CPU + 2 underpowered GPU's, they should go to 1 CPU and 1 GPU, then they truly don't get it.
The high end of the market is comfortably using 2P workstations (dual E5-2687w v4 or better) with 2-4 high end GPU's with 80+ lanes of PCIe, and even Thunderbolt if you want it.
That is the price of entry at the high end.
I hope this is more context around "Polaris is the plan, no wait Vega is the plan, aw crap."
I don't make the plans, I just hear about them. But I hope Apple's comments give more insight into that churn.
Definitely was not, as some people suggested, an attempt to spin down the Mac Pro or kill it.
Sure, if you want to use SATA. The Samsung NVME M.2 drives that the iMacs use retail for about five times that, which is why Apple charges $600+ for the upgrade.
Same. It doesn't mean I'm not (carefully) optimistic but I can see how this could be the sign of a company that was actually completely unaware of how important the pro user base is and how little patience many have left. I know some people who have left and are never coming back. I fortunately have a year left before I have to seriously consider an upgrade.
We may all end up seriously disappointed but this is going to be an exciting year nonetheless.
First to all the apologists for the trashcan mac pro here, I'd like extend my sincere HAhahahahahahhahahhahahhaaaaa (and worse) in your face. Even apple acknowledges (in an insultingly backhanded way devoid of any humility--aka the apple norm) what a train wreck it is. Thanks for prolonging this agony by being apologists for it for so long. I thoroughly expect you to now try to further justify it and will enjoy your monkey dance to do so.
This is a total strategic business failure. All the more shocking that they didn't learn from the cube debacle.
Apple's price is at record high, but my confidence in them is at record lows.
This is a total strategic business failure. All the more shocking that they didn't learn from the cube debacle.
Apple's price is at record high, but my confidence in them is at record lows.
Whoops, my mistake. Still, when you're a company that wants to offer the "best of the best" skimping on drives is silly.
Custom designs doesn't negative having flexibility in design tolerances. It isn't like Apple is building custom stuff for a deep space out planents probe with a fixed thermal/electrical budget of a highly restricted power supply. This is a desktop that plugs into the wall and stays plugged in all the time.
If the target is 150W GPU/CPU and put a +/- 30W tolerance on those then for 3 you have +/- 90W on the enclosure so that have some slack to play with if the vendor targets are slightly off. They didn't need to double the power and dimensions..... just some some design buffer build in.
All the more true if wave hands and claim you can't predict. If your estimates from the lab prototoypes typically runing with a 20% error bars then put the fracking error bars in the design. That is just basic reasonable engineering. Nobody designs a bridge brittlely exactly for some specific set of cars. There is an estimate. Then put a safety margin on it and built it.
I'm not sure who was smoking something that thought Vega was going to come in "small" on the TDP front. That has to have been clear for about a year.
I think you lost track that the RX 480 being appropriate was about the iMac ... not the Mac Pro. Apple isn't anywhere close to putting something like the 1080 or 1080 Ti in the iMac this year or in the future. If the 27" iMac simply move just to desktop CPUs along with desktop GPUs (versus mobile ones) that would be a substantive move forward. Moving to the top end of the mainstream GPU spectrum isn't really an issue.
As for the Mac Pro. two RX 480 class custom cards clearly would be better than the D300 they just had to drop. So if Apple had them they could have used them now. Gazing at the top end only is yet another bogus engineering move. Apple needed a Plan B because that exactly were they are now. If they were dogmatically committed to that exact same limited container restraints that Polaris 10 works.
Even if no revolutionary leap in performance still could get off the GCN 1.0 still stuck on now. Part of the Mac Pro problem was software adoption of new GPGPU related tech. Stuck on GCN 1.0 is an major impediment. Apple really can't move to OpenCL 2.0 even if they wanted to. If they are shifting to some other solutions they still aren't being helped by stick with the restrictions of GCN 1.0.
A RX 480 with 8GB is likely close to the D700 performance in alot of areas. Apple has nothing better than a D700 now. Apple is stuck with these GPUs for probably another year (at least. pretty good chance it will be more than year). If not going to change the rest of the system at all they are a better option than doing nothing. At the least the more adaptive, dynamic thermal control could survive better in the "too small" restraints they'd be stuck inside of. No, they are a multiple year solution but at the moment what Apple really needs is a "bit more than year" like solution. .... that they don't have.
If you read Apple's comments on it, without saying "Vega" they basically directly admit this is what happened.
Yes, maybe they were hoping the computer fairy was going to visit and make the Mac Pro's thermal tolerance better or Vega smaller.
But they directly said today they were trying different GPUs without success.
Should they have realized it earlier? Probably. But they also have AMD doing custom work for them and I think were hoping AMD could pull a rabbit out of a hat.
No. The 12 core is gone also.
Again this hints toward Apple long term plan is to go with the "workstation" socket that Intel has lined up. Skylake-W rather than a long term switch to the server targeted socket. The core range would be in the 4-10 range with higher clocks rather than chasing the "core count at whatever cost" path the server socket is shifting too.
Plus those 12 core + D700 were being largely picked up by folks who tended to drive them into a corner case that the current Mac Pro doesn't handle so well and don't expect to sell alot more of them. They have a smaller BOM range and use that to offset the cost reductions.
The folks who are in the "12 x86 cores really aren't enough camp". ... there is almost nothing in what Apple said in these sessions that points to them trying to chase after that niche.... even with the upgrade coming.
Yeah, but how many people using Mac Pros (that is the forum we're in), go more than 1TB. My iTunes library is pushing 4TB, and I still haven't moved a lot of my TV shows to it.
I guess it will be soft-unveiled this Fall along with hard-release of iMacs like the 2013 nMP was revealed at WWDC 2013 4 full months before release.
They have now freezed purchase for some customers by announcing a new model so I guess actual release can't be that far (first quarter 2018?).
Starting tomorrow morning they can be a lot less secretive with suppliers of the modular-MP (mMP?) parts now that the news are out..
As for SATA, if iMacs have SATA, the new mMP will have it as well...nowadays 2.5" drives go up to 5TB...could easily accomodate 4 of those once you're building the case around a full sized GPU and top of the line xeon anyway...they're in a no-BS mindset for the macpro and they wouldn't risk screwing up once again...and, I repeat, allowing HDDs in iMacs and not in MP would make zero sense given the no-BS mindset..
They have now freezed purchase for some customers by announcing a new model so I guess actual release can't be that far (first quarter 2018?).
Starting tomorrow morning they can be a lot less secretive with suppliers of the modular-MP (mMP?) parts now that the news are out..
As for SATA, if iMacs have SATA, the new mMP will have it as well...nowadays 2.5" drives go up to 5TB...could easily accomodate 4 of those once you're building the case around a full sized GPU and top of the line xeon anyway...they're in a no-BS mindset for the macpro and they wouldn't risk screwing up once again...and, I repeat, allowing HDDs in iMacs and not in MP would make zero sense given the no-BS mindset..
Yes, Gruber has updated his Daring Fireball post to note it's D700s on the $3999 model and not D800s. MacRumors and 9to5Mac have done / are doing the same.
SATA is already out in the Mac Minis, MacBooks, MacBook Pros... iMac is likely next. I wouldn't be surprised if this Fall they do away with hard drives in the high end models, and no more Fusion Drive.