MP All Models 40 core ASI with PCI-E slot is around the corner!!

[jmho]

I honestly think that it's in Apple's best interest to not release the Mac Pro at WWDC.

The only way their VR goggles aren't a complete unmitigated disaster is if Apple can hype up enough developers to burn $3k on them and make some software that might actually give normal people an excuse to buy them when they come with the consumer version.

There's also the fact that we're past the point of developers needing the absolute fastest machine money can buy, and 90% of coders are fine with a MacBook and aren't actually interested in a Mac Pro (even if they might want one)

I think it makes way more sense for Apple to drop the Mac Pro this month, give devs 2 months to geek out about it and then inevitably decide they can't justify the expense, move on, and then make sure the attention is 100% on the goggles when the WWDC keynote hits.

 

[ZombiePhysicist]

I honestly think that it's in Apple's best interest to not release the Mac Pro at WWDC.

The only way their VR goggles aren't a complete unmitigated disaster is if Apple can hype up enough developers to burn $3k on them and make some software that might actually give normal people an excuse to buy them when they come with the consumer version.

There's also the fact that we're past the point of developers needing the absolute fastest machine money can buy, and 90% of coders are fine with a MacBook and aren't actually interested in a Mac Pro (even if they might want one)

I think it makes way more sense for Apple to drop the Mac Pro this month, give devs 2 months to geek out about it and then inevitably decide they can't justify the expense, move on, and then make sure the attention is 100% on the goggles when the WWDC keynote hits.

We can agree to disagree. To me the idea of it being distracting is completely off.

However, the idea of the show being too long is fair. Depends if they have real meat on the bones in showing new features for macOS and iOS, or if they’re showing new emojis for 40minutes. If the rest of the show is full of boring stuff, a Mac Pro would be welcome in not causing them to stretch boring parts out. If it is a super packed show, then yes, moving the Mac Pro could make sense.

 

[Apple Knowledge Navigator]

Yeah, I agree with what’s been said. If the AR goggles are making an appearance at WWDC then Apple will want it to retain the spotlight.

Releasing the Mac Pro beforehand will not only provide the fastest developer platform, but will allow for enough PR time between it and the goggles.

 

[Mago]

It seems weird to me that they would release a Mac Pro before WWDC since they just announced the show date.

Agree, while WWDC not a purist software-only keynote, makes sense not to focus on new actually not revolutionary hardware (we should admit the Mac Pro doesn't revolutionize anything's, actually this role have always been for a mini or a MBP).

I honestly think that it's in Apple's best interest to not release the Mac Pro at WWDC.

The only way their VR goggles aren't a complete unmitigated disaster is if Apple can hype up enough developers to burn $3k on them and make some software that might actually give normal people an excuse to buy them when they come with the consumer version.

There's also the fact that we're past the point of developers needing the absolute fastest machine money can buy, and 90% of coders are fine with a MacBook and aren't actually interested in a Mac Pro (even if they might want one)

I agree, but in Cupertino there's no trust on AR/VR glasses, but AI development suddenly becoming top priority for apple, is where the Mac Pro means an solid proposal Worth to watch.

Releasing the Mac Pro beforehand will not only provide the fastest developer platform, but will allow for enough PR time between it and the goggles.

Again, assuming VR/AR glasses still hype at Cupertino.

---

I'm not confident neither I'll do bets on AR/VR technology, first and foremost besides simulators it doesn't have an HALO application yet, the best concept are on gaming (where apple never been competitive), and at 3k$ without a clear Halo application I won't put money on it.

Meanwhile, ***GPT (AI) just rise it's momentum flag, Apple should be crazy now to avoid at all costs not being left behind, after SIRI losing its Halo years ago, now *GPT (particularly as MS is deploying It), will revolutionize every workflow from coders to novel writer's, doctors, financial analysts, designers etc.

Is where the Mac Pro is key Central piece (specially in Cheese grater form), but not as a simple quad m2 max luxury workstation, it needs That magic "one more thing": an powerful AI training accelerator peripheral.

It's where ASi may pay itself if Apple provides pure TPU add on cards to supercharge the Mac Pro AI capabilities.

Said TPU peripheral likely on advanced development now and may explain a Mac Pro release at WWDC.

Despite how Cool is the ASi Trashcan, the Mac Pro I really want is one where I can put 800W+ on tensor accelerators peripherals.

 

[Boil]

First of all, no thanks to a new Trashcan, I would rather see a new Cube...

Second, Apple needs to hurry up with their ar/vr goggles, I need to meet Hiro at the Black Sun...

Third, ;^p

 

[MacPoulet]

It seems weird to me that they would release a Mac Pro before WWDC since they just announced the show date. I guess crazier things have happened.

I guess if the machine really sucked a pr release could make sense not to draw attention to how much it sucks, maybe? But it otherwise seems ideal for WWDC release, where they could boast about it a bit.

Or it’s aimed at the video market and the release coincides with NAB…

 

[NC12]

First of all, no thanks to a new Trashcan, I would rather see a new Cube...

Agreed, people in this thread seem to be forgetting why the trashcan was a failure. The Mac Pro is supposed to be the top of the line modular machine. The trash can philosophy works but only for something like the Mac Studio. If they released the 2013 trash can as a Mac Studio while continuing to provide an upgraded Mac Pro with the 5,1 design Apple’s reputation in the pro market wouldn’t have been damaged

 

[goMac]

March made more sense than April or May. It’s getting too close to WWDC. And I don’t think the Mac Pro is going to get a quiet update.

The past two Mac Pro updates have been at WWDC.

Plus if there are any change with the Mac Pro developers will need to know about - those will be WWDC sessions.

I don’t have any insider information and could be wrong. But a release before WWDC doesn’t seem right to me.

 

[edanuff]

You need to be careful what you wish for. I can easily see an AI-themed Pro machine done in a way that pays very little dividends to people using the machine for video production purposes and in fact have a fair bit of suspicion that will be the way it goes. These compute modules might be solely designed for accelerating CoreML. WWDC would be absolutely the right place for such a release and the benchmarks would all be about showing how the Pro wins in PyTorch or Tensorflow benchmarks and showcasing generative AI apps.

 

[Mago]

The past two Mac Pro updates have been at WWDC.

Also I remember then wwdc keynote had lot of public in person, it's post pandemic apple the one that a year ago introduced the Mac studio with an online video , IMHO past Apple 'behavior' (did you realize apple is not an animal?) Don't rules how they Will do everything unless marketing and accounting says: it will pay for us. (I'm tired of Gurman-esque oversimplification, it has as much sense as my worst baseless prediction or 'leak')

These compute modules might be solely designed for accelerating CoreML.

Few facts:

There are two new "compute device" that run some special iOS, initially I believed it could be an "xeon-phi like" accelerator cards (Chinese ppl still believe apple is working on something alike), but not I got an solid tip on this just being Mac Studio booting an minimal iOS which would allow use it as a tethered device, even redshift tested its framework past WWDC, but also maybe an external e-gpu like specific device based on Mac studio with some disabled or defective cores, cheaper than an Mac studio but allows you borrow its compute into your mac-whichever.

Rumours: Apple is working on a compute accelerator, it's specific nature is unknown, maybe an xeon-phi like device on MXP modules populated wi Mx-max or Mx-Pro SOC, or an pure classic GPU on ASI. Also cooperation with AMD not paused maybe ASi Mac Pro would allow AMD GPUs as compute only accelerators.

My speculation, given AI prominence now, such compute accelerator should be tensor-matrix optimized without graphics specific cores.

Notwithstanding a ASi trashcan Mac Pro still possible, as it is ideal for video production (an field which Golden hour is now Behind us), also some serious but not revolutionary AI training, then apple may opt to offer cloud based AI training just like Google and MS, at some point it's the way as some projects simple can't be trained with few TPUs often hundred or thousand are used by weeks.

My cards the future Mac pro should be either one of these:

1. An cheesegrater like chassis with some revisions with m2 ultra and m2 extreme plus room for 2 MPX modules for compute accelerator peripherals and some PCIe slots for storage interface etc, 4tb ram support.
2. Or an ASi optimized trashcan with One or two M2-Ultra around a thermal core and 6-8 USB4 plus discreet RAM support for UpTo 4 TB .
3. And a 'One More thing': likely an ASi GPGPU solution and/or an ASi tensor training pure peripheral either as discreet PCIe slot or as an eGPU like gadget or as an Cloud service.

My evil inside bets are for ASi trashcan. + Cloud TPU training service, my hopeless fanboy inside bets for an cheese grater+ xeon-phi like M2 accelerator based on semi defective silicon mounted inside an MPX module.

 

[Romain_H]

Agreed, people in this thread seem to be forgetting why the trashcan was a failure.

Yes. But in how far is the Mac Studio any better? Ok, it doesn't overheat, but it is similar in it isn't upgradeable.
Functionally they are very similar. So Apple did not necessarily learn the lesson - or came to different conclusions on what the computing of the future looks like

 

[Adult80HD]

Yes. But in how far is the Mac Studio any better? Ok, it doesn't overheat, but it is similar in it isn't upgraeable.
Functionally they are very similar. So Apple did not necessarily learn the lesson - or came to different conclusions on what the computing of the future looks like

The Studio isn't the Mac Pro. It's that simple.

 

[Romain_H]

The Studio isn't the Mac Pro. It's that simple.

Maybe. Maybe not. Something like the studio didn't exist back then.

 

[Adult80HD]

Maybe. Maybe not. Something like the studio didn't exist back then.

Which is 100% the point. Apple was pretty clear that this was a new model, with a new name. It's NOT THE PRO. Period, full stop. Arguably it's the replacement for the trashcan, which was pretty much not upgradeable outside of RAM, and then not easily.

 

[Romain_H]

Pro, Studio, those are just names...

 

[leman]

Yes. But in how far is the Mac Studio any better? Ok, it doesn't overheat, but it is similar in it isn't upgradeable.
Functionally they are very similar. So Apple did not necessarily learn the lesson - or came to different conclusions on what the computing of the future looks like

Looking at their product timeline, it would seem that the conclusion they came to is that there is need for two products. A compact professional workstation with good performance, but limited expandability, and a large tower system for high-end applications and internal expandability. At first they tried to abandon the second product type all together (thus the cylinder MP was born), but then it became obvious that not everyone was happy with it. So they reintroduced the cylinder MP concept in the form of the iMac Pro (and later the Studio), while working on reintroducing a more premium tower solution.

There is no reason why these products can't coexist.

 

[theluggage]

Yes. But in how far is the Mac Studio any better?

(a) Its significantly cheaper. Starting at $2000 or $4000 for the Max (10 core) and Ultra (20 core) versions vs. $3000 or $6500 for the 6 (originally 4) and 12 core trashcans.

(b) TB3 > TB2 and there's a much better range of Thunderbolt and USB 3.1g2 peripherals available now.

(c) It didn't replace the Mac Pro - the 2019 Mac Pro is still a viable system for Mac users who need massive internal expansion and high-end GPUs and probably has a couple of years left in it, even with Intel. One problem with the trashcan was that by the time it arrived, the previous Mac Pro was long overdue a meaningful update and had actually been discontinued for some time in Europe (because Apple couldn't be bothered to fit a fan guard to meet some safety regs that had been announced years previously).

(d) It had a very specialised triangular thermal design that relied on the heat emission being spread 3 ways between a single CPU and two GPUs and was essentially built around the specific Intel and AMD chips available at the time. ...the main lesson Apple admitted to was not being able to update it to newer CPUs or GPUs after 4 years because of this mistake. Of course, the Studio is equally dependent on the idea of a single, powerful system-on-a-chip - we'll see how that works out - but at least its now up to Apple to produce compatible chips.

...but, also, ten years have passed since the Trashcan vs Cheese Grater argument started, and the super-powerful personal workstation market that the 2010 and 2019 addressed is being nibbled away by faster laptops and SFF systems at one end and power-on-demand cloud computing at the other.

 

[Boil]

My hopes for an ASi Mac Pro...

• A17/3nm-based M3 Ultra ("standard" 2-way UltraFusion) & M3 Extreme ("Redfern" 4-way UltraFusion) options
• LPDDR5X RAM on both sides of M3 Extreme package (makes up for losing two edges of individual dies to UF connections)
• Symmetrical ("regular" die pairings) & Asymmetrical ("regular" & "GPU-specific" die pairings) SoC configuration options
• Hardware ray-tracing
• Cheesegrater 2.0 chassis for those who need PCIe slots
• New Cube chassis for those who do not need PCIe slots (M3 Extreme option only)

 

[Boil]

M3 Ultra

• 32-core CPU (24P/8E)
• 96-core GPU (w/hardware ray-tracing)

M3 Ultra-X

• 16-core CPU (12P/4E)
• 144-core GPU (w/hardware ray-tracing)

M3 Extreme

• 64-core CPU (48P/16E)
• 192-core GPU (w/hardware ray-tracing)

M3 Extreme-X

• 32-core CPU (24P/8E)
• 288-core GPU (w/hardware ray-tracing)

 

[Adult80HD]

My hopes for an ASi Mac Pro...

• A17/3nm-based M3 Ultra ("standard" 2-way UltraFusion) & M3 Extreme ("Redfern" 4-way UltraFusion) options
• LPDDR5X RAM on both sides of M3 Extreme package (makes up for losing two edges of individual dies to UF connections)
• Symmetrical ("regular" die pairings) & Asymmetrical ("regular" & "GPU-specific" die pairings) SoC configuration options
• Hardware ray-tracing
• Cheesegrater 2.0 chassis for those who need PCIe slots
• New Cube chassis for those who do not need PCIe slots (M3 Extreme option only)

I think the release will cover your 1st, 4th and 5th points. Second point is a maybe, the others highly unlikely.

 

[deconstruct60]

Yes. But in how far is the Mac Studio any better? Ok, it doesn't overheat, but it is similar in it isn't upgradeable.
Functionally they are very similar. So Apple did not necessarily learn the lesson - or came to different conclusions on what the computing of the future looks like

The 'lesson' that Apple learned was that it mostly works as a product, but do not explicitly label it 'Mac Pro'. The iMac Pro worked as a product. The Studio is working as a product. There are a bunch of folks who try to proclaim the MP 2013 as a complete failure, but MacStadium was buying pallets of MP 2013 systems right up until the last day Apple shipped them. Some folks did find use and utility for the system. Substantially unlikely Apple lost money on the MP 2013 as a product line. [ even more so given they snipped off macOS support relatively quickly after stopped selling the model. ]

Was the MP 2013 a 'license to print profits' money winner for Apple? Again substantially unlikely, but 'failure' is also a stretch. The overall Mac ecosystem grew (in both profits and units sold) over the product lifespan of the MP 2013. The MP 2013 lost some share as some folks moved to Windows and others did extended boycotts. But did that cause Apple substantive financial distress? No.

When Apple introduced the Mac Studio they went out of their way to say the Mac Pro was going to be something different (and that it replaced the iMac 27" model ) . Essentially, they did a very similar thing back in April 2017 (for iMac Pro. doing this Pro iMac now .... Mac Pro is something different substantively later. ) ,

It is far more so about Apple adjusting the 'scope' of the Mac Pro for the computing of the future. The MP 2019 raised the entry price 100%. For a sizable number of the old 2006-2010 crowd that was pragmatically a product shift change. The T2 brought a non commodity primary boot drive. SATA and alternative NVMe drives where options (but not BTO options from Apple).

Overall Personal Computing from 2004-2006 era to the 2020-2023 era has changed. Went from Desktops being dominate lead technology drivers and market makers to laptops playing that role. One of few narrow niches that hasn't shifted is toward the upper 25 percentile of the class "Mac Pro" space. Where lean on multiple internal drives, larger RAM capacities , and much higher than average internal bandwidth. That space isn't sole definition of "future computing" .

Apple needs something with more internal storage capacity than a Studio to creditably play in the Mac Pro space. A single internal drive too limiting. For alternative fast SSD storage that pragmatically means some sort of internal PCI-e slots. Apple probably isn't going to build some temple to hyper-modularity on every aspect possible, but also pretty likely won't be 'repeating' the exact same constraints as the Mac Studio on another product. (they already have a Mac Studio in the line up. Two helps them how? )

 

[ZombiePhysicist]

Trashcan was a failure on multiple levels. Anything that forces an apology tour is a failure. That it worked for some people is as irrelevant as are the people that love their brown Zunes.

 

[AlphaCentauri]

Trashcan worked really well for many music studios. I had one myself. It should have never been called a replacement for Mac Pro, though. It should have been called “Mac Studio” or “Mac mini Pro Max Plus” 🤣😉

As we now have a modernised trashcan in a form of Mac Studio, I expect Mac Pro to be an entirely different beast.

 

[Adult80HD]

Trashcan was a failure on multiple levels. Anything that forces an apology tour is a failure. That it worked for some people is as irrelevant as are the people that love their brown Zunes.

I used to be a CIO at a couple of large organizations so I would get invited to special Microsoft CIO Summit events in Redmond where they would wine and dine us and try and get us to spend more money. One year one of the welcome gifts was a Zune. A BROWN Zune. They had to literally give those away. I still never even opened the box. It was a literal and figurative turd, right down to the color.

 

[edanuff]

Used to be at a somewhat large tech company, around 120K employees at the time, and the trashcan was the standard (and only) desktop Mac you could get. Of course, most used MBPs rather than desktops and a lot of the desktop users were using Linux PCs, but there were a whole lot of trashcan Macs too.