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

Is the Mac Pro 7,1 a hit or a miss?

  • Yes

    Votes: 47 46.1%
  • No

    Votes: 24 23.5%
  • Not sure

    Votes: 31 30.4%

  • Total voters
    102
I think it's still too early to tell one way or another. If the MP 8,1 has some lower entry price points either with Intel, or ARM chips, that would go along way into saying it's a hit. If they continue to exclude semi-pros that don't need the top specs, but still want the design and expandability it offers, then I'd say it's a miss. Missed opportunities from Apple shouldn't come as a shocker.
 
  • Like
Reactions: JazzyGB1
I've had my 7,1 for a bit over a year and use it for my work and personal use, it is nice and I enjoy it a lot. I bought a bare bones one and upgraded memory to 96GB and put in an external 1TB system drive. I have been stuck at Catalina because I have an older Drobo and if I upgrade it will not work anymore... Drobo still has no new stuff in stock yet thanks to Covid stuff...
 
  • Like
Reactions: OkiRun
put in an external 1TB system drive.
What does that mean❓ Is the drive external or did you install internally (put in)
Unknown.jpeg


Lou
 
  • Haha
Reactions: OkiRun
I consider it a miss, strictly in my opinion only, because it did not persuade me to buy one. Technology has moved on tremendously since the 7,1 came out, and Apple needs a new flagship machine.
 
A hit - expandability, updated hardware specifications and x86 VM support.

Personal point - I HATE Windows (total rubbish) even though I am a software developer.
 
I bought a 16 core machine back in December 2019 [...].

Long term, however, remains to be seen. If Mr. Kuo's rumors about the machine and its successor are correct, then it will have been a wise choice as it will almost certainly be supported for years to come by Apple. Big Navi support is absolutely required long-term for the platform, and **if** there's another Intel version in the wings as he alluded to it probably means Big Navi support is coming. As long as it can support development on the current version of Mac OS/iOS it will do all I need. If it lasts me a decade as a workable machine it will have been a big "hit". I doubt it will, but 7 years is certainly possible.
Big Navi support is in 11.4. This, if nothing else, means we haven't been abandoned (yet). Also, I suspect this means we'll get MPX cards with Big Navi -- WWDC maybe? Regardless, it shows Apple is still actively supporting the Mac Pro 7,1.
 
^^^^You can eject internal drives too. AFAIK the operation of a disk labeled external or internal is the same. I have ten SSDs mounted internally. Six are listed as external and four as internal. Makes no nevermind to me.

Lou
 
Not many people put the specifications of the AS Mac Pro Mini they need in order to spend $3,500 usd on it; what does the 8,1 need to have for this amount of money? New Mac Pro Mini might be miss for some people.

If Apple doesn't give this specs, then those waited for time after the 7,1 will have to go to AS iMac or decide no Apple.
Or buy used 7,1.

It doesn't look like Mac Pro Mini will sell for less than 3k, base price. 5.5k for very top price. This is my experience with Apple pricing.

If Mac Pro Mini is a miss, then will perception be that 7,1 is less of miss?
 
Rumor smells funny to me. I'd be very surprised if we ever see that many efficiency cores on a desktop. Especially nearly 20% efficiency cores.


It wouldn't be surprising if Apple were going to try to use the same SoC "Building block" for the whole rest of the line up. ( e.g. iPad Pro -> lower end mini + lower end laptops gets one design , MBP 14-16 , iMac , Mac Pro "short" , Mac Pro "tall" gets a more enabld building block. )

That would get rid of Apple's not enough volume to scale with the Mac Pro. If MBP 16" and iMac 27 (or 32") were soaking up most of the SoCs then it gets the volume to get green light. The Mac Pro just "makes do" with what is available.

So if have a basic 16 core building block ( 12 big, 4 small , GPU block , NPU block , etc. ) then


M_Tile1.png


It would be a major departure from the M1 but if had a "Unified Memory Inter-package" crosslink hub then can get to 32 cores with a just closely bonding two SoCs together on the main board. Mirror that dual cluster along the bottom edge and could have point to point one hop links to all three "neighbors.". Apple could use the high speed interconnect not using on the non "four corners" edge to link up a limited PCI-e v5 hub/switch for slots (which the vast majority of the rest of Mac line up won't need. ) or hook a in house 5G modem ( which iMac, "max" Mini or Mac Pro may not need. ) .

A short Mac Pro with 2-3 slots might get just one PCI-e hub/switch whereas a "tall" Mac Pro could get two (and still provision out 7-8 slots ).

Apple doesn't have to reinvent the wheel here. CXL over PCI-e v5 physical foundation could be the interpackage link. Or they could invent their own.


For the single , SoC building block set up it could have a decent of "dead" high speed links but Apple can make those users just "eat" the costs. Similar on other end of the spectrum. It would be advent of the real nightmare some folks have dreaded where buying more RAM means making folks buy more cores too ( and extra Secure enclaves not using).

Could have incrementally lower RAM costs with just two 16GB stacks per building block. ( e.g. 32 cores and just 64GB ). Pretty good chance Apple would loose more than a few folks in the > 256GB data working set footprint market , but I suspect they wouldn't loose sleep over that if it cleaned up overall development path.

[ The iMac 24" ethernet port is out in the power dongle now, because had to turn that imac into bonus sized iPad in terms of thickness. ]


The GPU cores spread out over 2-4 packages will probably be decent for "compute" but probably won't win any major frame rate and/or large mega resolution monitor battles with AMD/Nvidia's top offerings. Again I doubt Apple will be pressed about that. ( Even more so if eventually let AMD back in the door as an option on Mac Pro's and perhaps large screen iMacs. )


Apple would also need to make some substantiative additions to hardware performance management logic ( and/or OS kernel changes) to distribute/manage the loads over the four packages. But the lower the fixed max number of packages should make that tractable.





It's a big desktop. Why would you implement power efficient cores? And M1 efficiency cores seem even more underpowered than their mobile counterparts.

As more apps expect there to be be power efficient cores laying around, some will spawn off more work to them. If mostly have developers highly weaned on developing phone apps that number is even higher.

I haven't seen much evidence that these power efficient cores are any "weaker" than the other ones in A13/A14. To the contrary, core development is rumored to be all "hand me down " iPhone cores just packaged differently.
( The M1 is in the iPad Pro just as much as it is in the MBA ).

In short, not particularly funny if walk away from premise that Apple is going to build something specific just for the Mac Pro. if they are not then probably stuff more targeted at those other Macs is going to seep into the Mac Pro configurations. IMHO, minimally going to get iGPU cores in the mix. wouldn't be too surprising is the power efficient cores came along for the ride also.
 
Looking at the rumor again, the other smelly thing is where it implies that there will be 512 gigs of RAM on an SOC.
That seems... highly improbable.

RAM density is going up with DDR5. But it also doesn't have to be just one SoC ( if willing to take on some mild NUMA issues. ). It doesn't have to be just one, all encompassing , interposer to hold all the RAM.

Getting to 1-2TB of RAM ... yeah that probably would turn into a problem. There is a decent chance though Apple won't pursue that capability.







Gurman has already weighed in that there is no full size Apple Silicon Mac Pro in design right now. I'm more inclined to go with that. Maybe the above rumor is talking about the Mini, but at those prices and configurations doesn't seem like it.

Even for the "Short" ( about half sized ) Mac Pro going up to 512GB and 64 cores there would be plenty internal volume to keep that cooled and running.

The Mini is also likely not going to be highly enabled by Apple to fratricide out the large screen iMac.
They may let the Mini and the 24" battle each other more. But I would be quite surprised if they let the Mini leveled up to the top end large screen iMac.

Also not likely Apple is going to want to charge past 64 cores in a single system image either. ( kernel changes needs don't mesh up with the primary objectives with iOS , tvOS , etc or even more of the macOS user base. )
 
Last edited:
Big Navi support is in 11.4. This, if nothing else, means we haven't been abandoned (yet). Also, I suspect this means we'll get MPX cards with Big Navi -- WWDC maybe? Regardless, it shows Apple is still actively supporting the Mac Pro 7,1.

Big Navi Support doesn't necessarily mean MPX cards. Big Navi GPU chips are so scarce that even just "plain" cards would hard to ship in volume by WWDC.

Big Navi support on Intel (x86-64) opens door for eGPU usage also. Again, if the cryptomining were not sucking the life out of the availability market it isn't just a Mac Pro only feature.

Apple putting Big Navi in the macOS x86-64 branch of 11.4 would be a prudent precursor to putting Big Navi support into macOS 12. WWDC that really want need to look at. Not a new card. But any light on any progress in getting non-Apple GPU support into the ARM branch of macOS. Apple may want to give Apple GPU over a year "head start" before opening up GPU competition. ( bigger "hammer" to make more developers optimize heavily for their GPU specifics. )


The original Navi drivers were not spectacularly stable and smooth. Just getting a Navi2 version "0.x" into 11.4 doesn't mean going to have product smooth drivers.
 
As more apps expect there to be be power efficient cores laying around, some will spawn off more work to them. If mostly have developers highly weaned on developing phone apps that number is even higher.

Worth noting this isn't how power efficient cores work at all under Darwin/macOS/iOS.

An application actually can't see power efficient cores and performance cores. So developer can't design apps for them or expect them. The kernel will send work to the power efficient cores as it sees fit. And... that's the whole deal. It's all automagic. The app can't expect anything, or directly spawn work to power efficient cores. Not even on the phone.

Which takes me back to my original thing of the high number of power efficient cores making no sense.

There's some reasons why things are probably designed this way. Some Apple CPUs run in a power or performance core exclusive mode, and that could get messy for a developer to wrangle. Some Apple devices will automatically turn on and off certain cores and certain core types, so your performance or efficiency cores may just disappear mid operation.

Regardless, apps can't be designed around core types.

It can make threading difficult because the big/little Apple Silicon CPUs can create really imbalanced processing. But you just have to split your work into small enough chunks that if the kernel decides to split work on different core types it doesn't really throw things off.

I haven't seen much evidence that these power efficient cores are any "weaker" than the other ones in A13/A14. To the contrary, core development is rumored to be all "hand me down " iPhone cores just packaged differently.
( The M1 is in the iPad Pro just as much as it is in the MBA ).

They're fairly weak compared to a performance M1 core.

Getting to 1-2TB of RAM ... yeah that probably would turn into a problem. There is a decent chance though Apple won't pursue that capability.

Again, given that Gurman is indicating Apple is strongly considering keeping the full size Mac Pro on Intel for the time being, I think Apple may be fairly stuck. It seems like they do want to keep a tower around with 1-2 TB of RAM, and probably upgradability.

In a few years, as they work out more Apple Silicon designs that have a higher PCIe lane count, maybe more memory channels, they could probably swing back around to the full size Mac Pro. But from the rumor mill whispers it does really feel like there is no grand master plan right now for the full size Mac Pro, beyond that it needs to exist.

(I'm sorry I didn't reply to more of your comments individually because I think most things really go back to ^. Apple recognizes the need for the role the full size Mac Pro fills but has no Apple Silicon solution on the near term roadmap for it.)
 
Last edited:
I've had the MP since early 2020 and it has been the most stable Mac I've ever owned. I don't think it has crashed one single time. Literally the only reason it ever gets restarted is to update it, and right now it has over a month of uptime. I don't have any complaints about it!

My 2019 16" MBP on the other hand has been a total piece of ****, and the internals have already been replaced once by Apple lol

Screen Shot 2021-04-22 at 12.02.04 PM.png
 
I've had the MP since early 2020 and it has been the most stable Mac I've ever owned. I don't think it has crashed one single time. Literally the only reason it ever gets restarted is to update it, and right now it has over a month of uptime. I don't have any complaints about it!

My 2019 16" MBP on the other hand has been a total piece of ****, and the internals have already been replaced once by Apple lol

View attachment 1762197
That is incredible.
 
That is incredible.
What is - uptime of 43 days?
I had like 200 days on my Trashcan until a weird and so far very rare Wacom driver bug caused a panic on wakeup. The counter is at 88 days again now. My mini on the other hand is at 460 or so days of uptime. Sleep function - it's convenient and it works! :)
 
I have a 16 core Mac Pro 7,1. That said, I consider Apple's CPU technology choice to be a miss. The platform would have been better served with the AMD Epyc CPUs. AMD has been aggressively improving the performance of all facets of their lineup from consumer to prosumer (threadripper) to workstation/server (Epyc). Better yet, they have retained the socket for multiple generations, allowing for a reasonable upgrade path.

I have seen many comparisons of the Threadripper (for example the AMD 3970X), touted as a Mac Pro killer. Unfortunately, while the threadripper is fast on paper (and indeed performs as expected for some applications), fails miserably for memory intensive multi-threaded applications. This is due to the limited memory bandwidth and the architecture of the chiplet design.

Before buying the Mac Pro, I built a 3970X based system and did some benchmarking with real-world multithreaded applications (scientific / numerical work I do). Shockingly, the 3970X performed very poorly on these applications and did not live up to the expected benefit of 32 core / 64 threads. I did expect some degradation (i.e. non-linear speedup), but what I experienced was extremely poor.

For CPU bound, light memory applications, such as raytracing and rendering, the 3970X performed as expected. However for multithreaded memory intensive applications I run, performed worse than my old 6 thread i7-8700K processor. I returned the system and decided to give the Mac Pro a try. Fortunately the Xeon W 16 core cpu performed well.

Are the Intel Xeon CPUs worth the cost?, not really. Epyc has similar or better memory bandwidth and more cores for a similar or cheaper price.

Let's hope that there are some CPU upgrades in the pipeline at Intel for the 3647 socket, providing additional performance for our Mac Pro 7,1 investment.
 
  • Like
Reactions: OkiRun
What is - uptime of 43 days?
I had like 200 days on my Trashcan until a weird and so far very rare Wacom driver bug caused a panic on wakeup. The counter is at 88 days again now. My mini on the other hand is at 460 or so days of uptime. Sleep function - it's convenient and it works! :)
Yeah. The up time.

200 days is crazy. I'm just going to leave my macs on and see what happens.
 
Yeah. The up time.

200 days is crazy. I'm just going to leave my macs on and see what happens.

It should just work, really.

20+ years ago as a student I helped run a computer lab at uni, filled with unix workstations. The students d*cked around on them all day running self compiled code and whatnot but we restarted the machines only over the christmas holidays. It was very rare that you had to actually reboot one because of a problem. That I found amazing.

Assuming the Mac Pro is an overall stable system it should be able to get there easy if only used by one or very few responsible adults. :)
 
  • Haha
Reactions: OkiRun
I would love to invest in an 8,1. My disappointment will be if the 8,1 would be as easily configurable as the 7,1 with off-the-shelf parts. Hoping this isn't the case and they move along the lines of user replaceability. If not, that would be a hard sell for people like myself who have a 7,1 and love tinkering.
 
The 7,1 was/is completely underwhelming and almost as much of a misstep as the 6,1. It's just not a great computer for a lot of tasks and too specialised, especially considering the hype ("we listened to and worked with creative professionals to make this machine" - shame you didn't talk to anyone using Adobe CC) and the promises of GPU rendering which are still mostly unfulfilled. The Xeons are uninspiring and return woeful clock speeds on single core tasks, GPU options are limited and beyond that, there's nothing really to get excited about.

The 8,1 needs to be an absolute powerhouse of Apple Silicon including CPU and GPU performance. No more compromises, empty hype or excuses. Just give us thirsty pros a machine to love and be inspired by again. We have been patient and loyal and need to upgrade our ancient 5,1 more than ever.

A less wallet gouging display would be nice as well.
 
The 7,1 is perfect for my needs. I can build a faster machine and equip with multiple displays for the same price as the MacPro + XDR combination, but I chose performance + industrial design and engineering. I don't know any other desktop with this kind of build and silence. No loud fans, water cooling, or dancing lights.

They could have gone with AMD processors and a PCIe 4.0 but they knew they were making a big switch. Plus a faster processor is only an excuse for poor software optimization of the actual platform they are coding to. Every major Apple software developer had a chance to optimize their software. It's not Apple's fault Adobe is still stuck on their archaic platform. While Black Magic's DaVinci chose to optimize theirs. Adobe is well aware of their lack in performance and they chose to move forward with Apple Silicon optimization versus rewriting their current platform for Intel.
 
The 7,1 was/is completely underwhelming and almost as much of a misstep as the 6,1. It's just not a great computer for a lot of tasks and too specialised, especially considering the hype ("we listened to and worked with creative professionals to make this machine" - shame you didn't talk to anyone using Adobe CC) and the promises of GPU rendering which are still mostly unfulfilled. The Xeons are uninspiring and return woeful clock speeds on single core tasks, GPU options are limited and beyond that, there's nothing really to get excited about.

The 8,1 needs to be an absolute powerhouse of Apple Silicon including CPU and GPU performance. No more compromises, empty hype or excuses. Just give us thirsty pros a machine to love and be inspired by again. We have been patient and loyal and need to upgrade our ancient 5,1 more than ever.

A less wallet gouging display would be nice as well.

I think if the only options we have for GPU with apple silicon are apple GPUs, it will be a huge misfire. It still needs slots and be able to use standard PCIe cards from the PC world. That is key. I dont know how easy that is or not though.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.