MP 7,1 So...Who Ordered a New Mac Pro?

[rondocap]

I am curious who here ordered the new Mac Pro, or the Mac Studio - or decided against either.

I am a current owner of the 7.1 Mac Pro which I am still keeping and using, but I ordered a Mac Studio.

I was *This* close to ordering the new Mac Pro, but after seeing it's the exact same thing as the Mac Studio aside from the PCIe slots, I decided against it for the extra 3k.

If it had just like one more key point as a difference to the Mac Studio, I would have considered it more.

If I didn't already have a 7.1, I may have ordered the Mac Pro - just because I think the design and case are amazing, and I can make some use of the PCIe slots.

Pegasus R4i is an MPX module, so not supported on the new Mac Pro, which is a bit of a head scratcher imo. MPX was such an expensive and short lived adventure..

 

[flat4]

i'm waiting until the 7,1 is at 3,1 prices

 

[TrevorR90]

I'm not, going to stick with my 7,1 until it dies. I need something that I can stick my Nvidia gpus in being a data scientist. Also something that I can run Windows on.

Nvidia GPUs and Windows doesn't seem to be happening with Macs anymore so the 7,1 will be the last tower desktop from Apple for me.

I would consider a Mac studio or macbook in the future though.

 

[Foxdog175]

Coming from a maxed out 6,1, I was hoping not to be tempted by the new Mac Pro, mostly due to the footprint. All of my audio hardware is outside the box, so I don't need PCIe either. They could have delayed until M3, or until they figured out an extreme chip...some reason to choose a MP over a Studio, but I'm happy to say I wasn't tempted in the least. Ordered a maxed out Studio without hesitation.

I feel for those who need PCIe though. It feels like Apple rushed a new Mac Pro just to complete the ASi transition.

 

[richinaus]

Gave up on the mac pro when the M1 was released as the writing was on the wall, and shifted to PC's to complete heavy lifting for rendering. Nvidia have the market here.

 

[singhs.apps]

The Mac Pro is a joke.

I would have gotten the studio if I hadn’t upgraded my PC workstation 8 months back.

Right now Mac chores are via a maxed out M1 Max MBP.

Future ?

Mac Pro only if it has extreme chips, better than 5090 or above GPU performance + discreet GPUs as expansion modules + 384 and above ram.

Else a M3/M4 ultra studio.

Or maybe not even that and call it a day from working in macOS.

 

[m1maverick]

I didn't order any new system as I currently own an M1 Studio Ultra and my Mac Pro needs are met with my Z840 system. The direction Apple is heading the Ultra may be my last Macintosh.

 

[PowerMike G5]

If I didn't own the 7,1, perhaps I'd maybe get one.

It'd be a touch decision. I just wish there was more differentiation between the 8,1 and the studio. I guess I'll see how I feel when the 9,1 comes out. With AS, I do think there will be a more frequent cadence of updates at least.

 

[MisterAndrew]

Even if you like the Mac Pro 8,1 I think it would be best to hold out until 9,1 or 10,1 since there should be quite a bit of improvement by then. This first AS Mac Pro is the guinea pig and was something Apple quickly hobbled together to complete the transition. The 7,1 will be supported for at least 3 more years which takes care of PCIe slot needs. A Studio can fill in for AS needs in the meantime, but doesn't offer much of an advantage over a 28 core Xeon with 6800 or 6900 AMD graphics.

 

[prefuse07]

One thing I wonder though -- if the 8,1 has really low sales numbers.... Will Apple just say "see look, nobody wants it" and scrap the product line altogether?

In other words, they purposefully handicapped it, to make it easy for them to justify scrapping it.

That leaves us with a pretty grim outlook, no?

What do you guys think?

 

[goMac]

One thing I wonder though -- if the 8,1 has really low sales numbers.... Will Apple just say "see look, nobody wants it" and scrap the product line altogether?

Very possible.

With the feeling that maybe development had a lot of issues - maybe they know it has issues and this is a stop gap. Maybe they're planning on fixing this with the next Mac Pro.

Or maybe this is going to be what it is, and it gets cut.

At least it shares its guts with the Mac Studio so it's probably not that expensive to design a new one. Just update the SOC, maybe the board, and you can keep going for a while.

 

[deconstruct60]

Pegasus R4i is an MPX module, so not supported on the new Mac Pro, which is a bit of a head scratcher imo. MPX was such an expensive and short lived adventure..

The MPX was an expensive adventure for customers or Apple? For Apple, the MPX connector ( in an of itself) probably was not relatively expensive at all (**) . Apple had jump some hoops to route mulitple Display Port streams through a switch and hither-and-yon all over the logic board to remote ports. But that became a solution in search of a problem when just move the Thunderbolt controllers onto the main die. Even Intel moved TB controllers onto the main die in the laptop models in 2020 before Apple even launched M1.

The R4i is a rather poor justification for the MPX connector. It doesn't use 3/4 of the connector pins present. It also likely doesn't use anywhere remotely near 150W let alone the full budget of the connector power. Unless had some older , power hungry 3.5 HDDs , 75W is a decent budget to run four drives and a small embedded controller SoC with some RAM. 75W comes off the basic bus for PCI-e. And the R4i soaking up the x16 PCI-e v3 bandwidth for just 4 HDD disk drives ... another gross waste of potential resources.


The J2i product worked fine with a couple of cables. It also was 'cloned' as a product by multiple folks. For example,

Fusion Flex J3i (2019 Mac Pro Drive Bracket) - SONNETTECH

3-drive mounting system for Mac Pro (2019). Install SATA 3.5-inch hard drives and 2.5-inch SSDs. Kit includes brackets, screws, and cables.

www.sonnettech.com

[ Folks who just wanted an internal "Time Machine Drive" ... the J2i worked fine. Apple made "internal TM" popular on previous Mac Pro's so there was demand that pulled in solutions from multiple vendors. ]


The R4i was cloned ( best I can find) by NOBODY. Relatively extremely straight forward engineering changes could have turned it into a 6-pin + bus , or maybe bus only, powered card with no real material differences in feature set. The R4i 'smells' a lot like the LG UltraFine 21.5"/27" products. Apple kind-of-sort-want wanted something but didn't want to make it themselves. So they go to a vendor and to some "back seat driving" of the specs. ( "make a monitor with one and only one input, no buttons , etc. ... Yeah nothing like the other stuff you sell." "Make a huge card that takes four 3.5 drives with now wires for power. Yeah, nothing like the other stuff you sell at all") . In exchange for the pushing the constraints on the design, Apple promises to buy a sizable block of product ( should help mostly/partial pay for R&D indirectly ) and vendor gets "most favored " status on the Apple Store. Lots of exposure in tech docs and sales material and Apple Salesfolks training sessions.

[ Note: nobody every cloned the LG UltraFines either with Apple's imposed design constraints. ]

The R4i was deeply born out of the Power Mac G5 and the carry over of the 4 drive bays (and a bit of Apple "real" RAID card). Apple felts they couldn't dislodge some folks from their 2008-2012 models unless the customers could pull those 4 HDDs in the trays and move them over.

The Pegasus M4 ( now R4 ) works across the whole Mac line up and isn't materially slower at all. Hence the R4i is 'covered' in at least two directions by just Promise products. J2i is way more affordable. M4 is way more flexible.
Throw in modern PCI-e cards with multiple U.2 or M.2 SSDs. deliberate large SSD cards. Upcoming E1/E3 EDSFF drives. etc.


** ASUS whipped together their own power only , "look ma no wires" connector for GPUs.

Asus Demos RTX 4070 GPU With No Power Connectors on BTF Motherboard

Replacing the 16-pin power connector with a proprietary slot.

www.tomshardware.com

(possibly after the 16-pin drama broke out). Probably didn't have a ginormous R&D budget to do that.

 

[richinaus]

Very possible.

With the feeling that maybe development had a lot of issues - maybe they know it has issues and this is a stop gap. Maybe they're planning on fixing this with the next Mac Pro.

Or maybe this is going to be what it is, and it gets cut.

At least it shares its guts with the Mac Studio so it's probably not that expensive to design a new one. Just update the SOC, maybe the board, and you can keep going for a while.

Wishful thinking is my thought on this.
Mac Pro is dead or you won’t see a new one for 3 years whilst the studio gets an annual update.
They probably released this latest model just because they said they would and didn’t want to back track.

 

[deconstruct60]

One thing I wonder though -- if the 8,1 has really low sales numbers.... Will Apple just say "see look, nobody wants it" and scrap the product line altogether?

Depends upon if Apple had intended to put a "bigger than Ultra" SoC in the system , but that fumbled that due to design problems.

May get a 'let's fix that with more deliberate chiplet design so it is actually cost effective" and they'll try again.

Apple has gone MP 2013 --> iMac Pro --> Mac Studio. All iterating around the same basic constraints of :
i. literal desktop in a 7" x 7" fooprint ( or less)
ii. 400-500W power budget
iii. limited modularity.


If the original 2019 plan was 'stop at Ultra" and the "bigger than Ultra" was only an internal crazy 'side' experiment then, yeah whatever happens in next 12-18 months probably determines fate. There is a couple of things that indicate that wasn't the plan.

a. the leak covered in the thread of "40 core ASI just around the corner". There were 40 core systems floating around. It just wasn't viable in the M1 and M2 forms they had choosen.


b. Apple ripped out 3 of the 4 8-pin AUX power connectors. Again pretty big indication that the CPU 'zone' was suppose to soak up lots more power than just a plain old Ultra. ( 300-400W range rather than about half of that. )
Apple also keep the 1400W power supply.

In other words, they purposefully handicapped it, to make it easy for them to justify scrapping it.

Extremely doubtful they skipped the 40 core version just to purposefully handicap it. More likely that SoC just didn't work right ( on some technical and economical reasons). Apple wanted to do the AirPower 3 device wireless charger... it just didn't work right. The 40 core version appears to have died before they got on stage and made a big deal about it ( like the normal Apple R&D secretive process).

The move in 2019 to crank up the entry price 100% to $6K probably had some notion of needing to 'reseat' the core target market in the Apple Silicon era. ( Intel and AMD components would drop out , but the SoC replacing them wouldn't be 'cheap'. ). Apple isn't looking to keep 100% of the old user base. They just new a large enough new set of customers ( a mix of some old Mac Pro users and some new folks ).

That leaves us with a pretty grim outlook, no?

For those looking for an HP/Dell/Lenovo clone that 'check boxes' all the features of those system with maximum modularity, it has been grim. Wasn't going to happen. The vast bulk of the Xeon W-3000/2000 R&D is all paid for by the Intel server business. Apple had to invent the MPX connector largely because there was no GPU in the CPU die. To leave that GPU-less was not Apple's call at all. It was the 'work around' that had to do because Intel made the call on the CPU package design.

There are some bandwidth throughput speed bumps coming to LDDR5/6 over next year or two that could allow Apple to put ECC coverage into their "poor man's" HBM implementation without a noticeable drop off in end-user visible throughput. TSMC N3 means they need to redesign the Max die anyway. There should be a decent track record if the Studio can support > 1M sales or not.

Four very large , oddly shaped monolithic deals isn't a good chiplet design. If Apple stops trying to hammer a round peg in a square hole and possibly drop down to just 3 side-by-side compute chiplets with I/O on either side that would probably work much better and be less expensive. ( just have decouple the Studio from slavishly using laptop chips. ) . But they'll need to spread that 'desktop oriented' chiplet solution over more products than the Mac Pro. Without expanding into the Studio sales it likely won't work economically.

The Extreme is starting to look more and more like "A Bridge Too Far". Apple got wound up trying to do the biggest possible thing ( "biggest airdrop ever . ..more audacious broad front ever ... fastest charge up a two lane road ever ... ) as oppose to the more tactically sound thing.

What do you guys think?

As for the fact that Mac Pro isn't backordered by months ( like launch of 2013 and 2019 ). The MBP 15" and Mac Studio are not either. Some of this is Apple getting the demand forecast 'right' . Some of this is macroeconomic conditions of lots of folks just not throwing money at tech like drunken sailors in a strip club.

Some of this "it is going to fail" is folks projecting "If Apple doesn't make exactly what I want then it should fail". That is likely wrong. ( heard similar stuff when Apple shifted away from XServe. )

 

[deconstruct60]

At least it shares its guts with the Mac Studio so it's probably not that expensive to design a new one. Just update the SOC, maybe the board, and you can keep going for a while.

This is key factor. If it is what is the cheapest possible path to the absolute fattest margins then they solely stay stuck on leveraging monolithic laptop dies that pay for themselves with other products ( Max MPB 14"/16"/Studio pay for the dies and only then looking for cheapest alternative augment. )

No real chiplets ... just cheapest kludge possible. But honestly I think the Studio is long term 'not so great' shape in that context too. Eventually after a few iterations just doing a plain Max and dropping the Ultra is cheaper too.

The could Scrooge McDuck their way all the way out of a upper end desktop line up. Just sell headless laptops.

 

[avro707]

i'm waiting until the 7,1 is at 3,1 prices

You’ll be waiting a while probably.

2023 MP might be the last one and maybe Apple will just kill the desktop computer business off completely and everyone can use mega-powered iPads.

Make it look fancy, make people throw it out and buy new one if they need to upgrade.

I’m holding on to the 7,1 and will upgrade it to the maximum.

 

[rondocap]

The MPX was an expensive adventure for customers or Apple? For Apple, the MPX connector ( in an of itself) probably was not relatively expensive at all (**) . Apple had jump some hoops to route mulitple Display Port streams through a switch and hither-and-yon all over the logic board to remote ports. But that became a solution in search of a problem when just move the Thunderbolt controllers onto the main die. Even Intel moved TB controllers onto the main die in the laptop models in 2020 before Apple even launched M1.

The R4i is a rather poor justification for the MPX connector. It doesn't use 3/4 of the connector pins present. It also likely doesn't use anywhere remotely near 150W let alone the full budget of the connector power. Unless had some older , power hungry 3.5 HDDs , 75W is a decent budget to run four drives and a small embedded controller SoC with some RAM. 75W comes off the basic bus for PCI-e. And the R4i soaking up the x16 PCI-e v3 bandwidth for just 4 HDD disk drives ... another gross waste of potential resources.


The J2i product worked fine with a couple of cables. It also was 'cloned' as a product by multiple folks. For example,

Fusion Flex J3i (2019 Mac Pro Drive Bracket) - SONNETTECH

3-drive mounting system for Mac Pro (2019). Install SATA 3.5-inch hard drives and 2.5-inch SSDs. Kit includes brackets, screws, and cables.

www.sonnettech.com

[ Folks who just wanted an internal "Time Machine Drive" ... the J2i worked fine. Apple made "internal TM" popular on previous Mac Pro's so there was demand that pulled in solutions from multiple vendors. ]


The R4i was cloned ( best I can find) by NOBODY. Relatively extremely straight forward engineering changes could have turned it into a 6-pin + bus , or maybe bus only, powered card with no real material differences in feature set. The R4i 'smells' a lot like the LG UltraFine 21.5"/27" products. Apple kind-of-sort-want wanted something but didn't want to make it themselves. So they go to a vendor and to some "back seat driving" of the specs. ( "make a monitor with one and only one input, no buttons , etc. ... Yeah nothing like the other stuff you sell." "Make a huge card that takes four 3.5 drives with now wires for power. Yeah, nothing like the other stuff you sell at all") . In exchange for the pushing the constraints on the design, Apple promises to buy a sizable block of product ( should help mostly/partial pay for R&D indirectly ) and vendor gets "most favored " status on the Apple Store. Lots of exposure in tech docs and sales material and Apple Salesfolks training sessions.

[ Note: nobody every cloned the LG UltraFines either with Apple's imposed design constraints. ]

The R4i was deeply born out of the Power Mac G5 and the carry over of the 4 drive bays (and a bit of Apple "real" RAID card). Apple felts they couldn't dislodge some folks from their 2008-2012 models unless the customers could pull those 4 HDDs in the trays and move them over.

The Pegasus M4 ( now R4 ) works across the whole Mac line up and isn't materially slower at all. Hence the R4i is 'covered' in at least two directions by just Promise products. J2i is way more affordable. M4 is way more flexible.
Throw in modern PCI-e cards with multiple U.2 or M.2 SSDs. deliberate large SSD cards. Upcoming E1/E3 EDSFF drives. etc.


** ASUS whipped together their own power only , "look ma no wires" connector for GPUs.

Asus Demos RTX 4070 GPU With No Power Connectors on BTF Motherboard

Replacing the 16-pin power connector with a proprietary slot.

www.tomshardware.com

(possibly after the 16-pin drama broke out). Probably didn't have a ginormous R&D budget to do that.

I am using the R4i right now, and I like it overall. It is a little noisy, but otherwise it allows me to have a large amount of relatively fast storage within the Mac Pro itself, minimizing clutter outside.

It does fit in nicely with the Mac Pro MPX modules - I just wish it was starting that could be passed along to the new Mac Pro, and sadly it can't because of the MPX design it seems.

 

[okkibs]

The new Mac Pro is likely just a stopgap measure until they figure out a way to bring it up to the capabilities of the older Intel model, i.e. supporting more memory. They had the redesigned Mac Pro housing already, and they had the tech from the Studio, so why not make it available for users that prefer this form factor or require the PCIe slots?

Apple knows very well that copying another Mac's hardware with its restrictions and performance (and adding slots that are missing GPU support!) isn't going to make for an exciting product. But they surely had the plan for where the Mac Pro will be heading well before releasing it gimped like this.

If I were Apple, I'd be designing the next ASi iteration with an option for DDR5 memory slots in mind, and I'd offer an upgrade service where customers bring in their Mac Pro 2019/2023 in for a mainboard replacement and a yearly mainboard-replacement "upgrade" path is established.

I am convinced Apple won't ask us to replace the entire Mac Pro, and I am also convinced they'll extract as much money from their customers as possible, so offering that mainboard upgrade service makes perfect sense.

At that point the distinction between Studio and Pro will be clear: At the very least the Pro will support double the memory and more, and I assume there will be a chip beyond the Ultra that will be Pro exclusive. Since the naming scheme is already at a limit (what could be better than Ultra?) they'll call it something like P1 (performance) which allows for another Mac Pro exclusive series with varying performance from the P1 Pro/Max to the P1 Ultra.

And then instead of being forced to follow the yearly M release cycle -it would look silly if the Studio got the lastest M release and the Mac Pro didn't- they can establish a separate, for example 3 year cycle, so that they got a couple years in between Mac Pro releases.

Perhaps there won't be any user-upgradeable memory at all and the P1 will just double the max memory config from the Studio. I am hoping for more memory configs but 384GiB+ would at least be better than the measly 192GiB we get now.

 

[avro707]

Perhaps there won't be any user-upgradeable memory at all and the P1 will just double the max memory config from the Studio. I am hoping for more memory configs but 384GiB+ would at least be better than the measly 192GiB we get now.

What may happen is that people hang on to 7,1 as long as possible then swap to high spec PC workstations which can be upgraded to much more powerful levels than the new Mac Pro.

The user base might leave Apple completely and they can just be left to make disposable laptops, phones and tablets.

 

[rpmurray]

Waiting for M3.

 

[MisterAndrew]

One thing I wonder though -- if the 8,1 has really low sales numbers.... Will Apple just say "see look, nobody wants it" and scrap the product line altogether?

In other words, they purposefully handicapped it, to make it easy for them to justify scrapping it.

That leaves us with a pretty grim outlook, no?

What do you guys think?

I was thinking this too. Apple isn't dumb. They intentionally released a Mac Pro they know has the exact same capability as the Studio. They know there are very few people willing to spend $3k for PCIe slots. It could be to drive Studio sales because the overpriced Mac Pro makes the Studio look like an excellent deal.

 

[avro707]

I was thinking this too. Apple isn't dumb. They intentionally released a Mac Pro they know has the exact same capability as the Studio. They know there are very few people willing to spend $3k for PCIe slots. It could be to drive Studio sales because the overpriced Mac Pro makes the Studio look like an excellent deal.

I find it hard to believe a company could be that stupid to spend $$$ to intentionally kill off a major product line and destroy its own reputation. Maybe I'm just not understanding the internal Apple religion.

 

[Grumply]

I know a lot of people are arguing for the value-equation of the Studio vs the new Mac Pro, but by the time you buy enough Thunderbolt-PCIe expansion boxes to add cards, local storage etc. to the Mac Studio, that $3k difference will be sucked up pretty quickly. And with the Mac Pro box, you'll have nice quiet fans.

PCIe expansion boxes generally have nasty cheap fans that make a racket. So it'll be much nice to have those peripherals tucked away inside the tower, and cooled calmly and quietly.

 

[Matty_TypeR]

TB-4 is limited in its speed, TB-5 will be double the speed, so any future Mac Pro with a M3 in 18 months to 2 years time might have TB-5 which means anyone buying this 8.1 mac pro will be out of date and not upgradable with in 2 years.

I dont see any advantage in buying into this new Mac pro to be out dated in 2 years time if not before. Its like a studio with PCIe for mass storage only.

I will update my 7.1 to 28 core when needed and add more Ram when needed, but will not buy into the closed in Mac eco system of the Mac pro 8.1 with just storage upgrade's available. And who's to say there wont be a TB-5 PCIe card available for the Mac pro 7.1 when its released. but i expect apple will block that like the 7900xtx update they could provide if they wanted to for existing Mac pro 7.1 owners who would like a 7900xtx.

I have the feeling that my next purchase will be a thread ripper based machine anyway, where i can use maybe a 5090 in 18 months time which will blow the M3 ultra away, ray tracing extreme performance awaits.

 

[rpmurray]

I'm still not understanding what this new Mac Pro offers that we couldn't have gotten with an updated Intel version.