[Frank777]

Wait, what?

[psmith2.0]

I don't think they're even making one.

They have to know it's a limited appeal, niche Mac, so why even bother?*




*Them talking, not me.

[chucker]

I still think they'll do M2 Mac Pro, M3 Mac Studio, M4 Mac Pro, etc.

[PB PM]

I suspect Apple may be tracking what the users of such machines want for upgrade cycles closely. All accounts point to movie studios being the big buyers of Mac Pros, so Apple will want to maximize development time, and only upgrade them when the types of users who really need them want to upgrade. I mean if you are dropping $10k+ on a video render/editing rig, it's unlikely you'll be upgrading every 2-3 years, unless you are doing a big upgrade in your video capture equipment to a new format.

[Frank777]

It would make more sense then for the Mac Pro to wait for the M3, and the Studio to get an M2 upgrade.

The problem is that the Studio can be upgraded anytime, whereas the Pro has to wait for the new chips to be available in the Fall.

I don't see how we wait another month for WWDC and still don't get anything announced in the desktop arena.
Apple hasn't updated any Macs at all since January.

I know this WWDC will be the coming out party for those glasses no one is asking for.
But is throwing out a 15-inch Air and saying "wait for the Fall" really a viable strategy?

My current thinking is the Studio will get a silent M2 update right after the "M3 Mac Pro" is teased at WWDC.

[dglow]

Originally Posted by Frank777 It would make more sense then for the Mac Pro to wait for the M3, and the Studio to get an M2 upgrade. The problem is that the Studio can be upgraded anytime, whereas the Pro has to wait for the new chips to be available in the Fall.

Well, the Studio could get the M2 Max at any point since it's already out in the MBPs. But the Studio will also need an M2 Ultra – and it would be unlike Apple to update the Studio lineup piecemeal.

Given the that AS Mac Pro is rumored to also be Ultra-based, then if the Pro is waiting for the M3 series it may be that Apple have chosen to skip the M2 Ultra altogether.

[Frank777]

We're 25 days until WWDC. I'm still hoping Apple has nixed the silly idea of heading into a recession with a crazy expensive set of 'glasses' no one is asking for, and will instead pivot to featuring the Mac.

The Mac Pro probably needs the M3 and will wait until November.
But as the economy heads south, the value proposition of the larger iMac will need to return.

If Apple debuted a new 27-inch with an M2 Pro, paired with a new matching 15-inch MacBook Air, Mac sales would instantly rebound.

[PB PM]

I doubt adding a more expensive units will cause a massive rebound in sales. An upgraded 24” might help, but honestly during COVID just about anyone who needed a new computer got one, the need isn’t there right now.

[chucker]

Mac sales are above pre-COVID levels, so I don't think a significant increase is realistic.

The iMac in total is 10% of sales if that, and the potential high-end model is surely less than half that. Especially now that the Mac Studio exists.

[Frank777]

We're pretty much 10 days out from WWDC and I still don't understand where we're headed.

The rumour mill says just a 15" MacBook Air and the Glasses, but that still doesn't make any sense to me.

WWDC is a developer event. The Glasses, and xrOS (or whatever it's called) is the latest developer platform.
I presume AR/VR development isn't going to be low-end. These Glasses aren't going to be cheap, even if $3K is a misdirect.

Aren't these AR/VR environments going to be built on the Mac?
We currently lack a Mac Pro, a Mac Studio on the latest chip, and a prosumer iMac.

Exactly what Pro machines are we going to use to develop these new AR/VR world building apps?

[PB PM]

If the AR/VR setup is designed to run on an iPhone/iPad, every Mac for sale today could run the development program without a hitch most likely.

[Frank777]

Originally Posted by PB PM If the AR/VR setup is designed to run on an iPhone/iPad, every Mac for sale today could run the development program without a hitch most likely.

True, a dedicated AR machine wouldn't need to be that beefy. But development houses, game builders and movie studios are likely simultaneously building a host of projects for iPhone, iPad, Film/Video, Mac, and now Glasses, that would benefit from using Pro machines in their workflow.

That means you need to have Pro machines to sell them. Apple isn't in this to sell a bunch of Mac Minis and MacBook Airs.

I don't see how Apple does this WWDC without showing the Mac Pro, even if it's not out till October.

[PB PM]

Apple is kind of limited with the M chips right now, all they can do is throw more cores at problems and that will only go so far til the hit the power limit of the current designs. There hasn't been a big bump in per core performance from M1 to M2 either, so unless M3 is a huge jump I just don't see a Mac Pro bringing anything special to the table for for most developers that a tricked out Studio cannot.

[709]

Correct me if I'm wrong, but I feel like adding more CPU/GPU cores is the entire point and advantage of the M-series chips. "We have these chips with xCPU and xGPU cores. Now we've put two or more of these chips next to each other, and we're calling it this". It feels kinda lego-like in its simplicity, for a non-chip-architecture-knowing guy like me.

To the larger point though, I feel like the 'Pro' moniker days are coming to an end. Everyone is a Pro now that we all have pro tools for dirt cheap. Back in the late 90's/early 00's Avid took a massive shit on Mac users - out of the blue - loyally using Media Composer for film editing. There wasn't really a price-comparable competitor at the time so eventually Premiere and Final Cut both pushed hard with 'Pro'. And Apple went hard with 'Pro' audio and DVD authoring and others. I can see the reasoning for all that, they were competing for high-end users. Now we have 'Pro' apps on our ostensibly weakest devices, comparatively, with Final Cut, Logic and Premiere all available on iOS.

Why go with 'Pro' now though? It's a needless classification.

I don't think there will be a Mac Pro, not in a way we're thinking about it anyway. It may be a custom tier that's only sold to select customers. That's not unprecedented in Apple kit. What that custom tier entails – maybe BYOMac? – I can only imagine.

If there is a new Mac Pro and its just old kit with new chips I will be massively disappointed.

[Kickaha]

People get weird about expansion slots. :/. Somewhere they went from practical kit to a fetish.

And yes, more cores is absolutely the right approach. Cranking up clock speed? Okay… a 50% clock speed improvement (like we’ve seen that in years) and you get how much improvement?

Now compare that to going from one core to eight.

The MOARHERTZES approach worked fine when chips were dumb as hell and all sorts of stupid silicon tricks had to be pulled out of the ass to try and squeeze out a bit more performance.

That’s CISC.

RISC, on the other hand, the principle behind ARM designs such as Apple Silicon, says keep the design simple, but streamlined and capable of inter core connectivity from day one… then start slapping those little bastards together.

That scales… assuming you have a programming model, language semantics, and compiler tool chain that supports it.

Apple just happens to. Gee, wonder how that happened.

[chucker]

Originally Posted by Kickaha And yes, more cores is absolutely the right approach. Cranking up clock speed? Okay… a 50% clock speed improvement (like we’ve seen that in years) and you get how much improvement? Now compare that to going from one core to eight.

Sure, but it'd be going from twenty cores (which the Studio already has) to forty. Or eighty. There's… diminishing returns at that point.

Originally Posted by Kickaha That scales… assuming you have a programming model, language semantics, and compiler tool chain that supports it.

IOW, it barely does.

[PB PM]

Throwing more cores at it doesn’t work well for everything though, some things just need brute force of a single powerful core. Does throwing more cores work well for some things? Absolutely. 3D rendering (GPU), yes. Video rendering? Yes (GPU). Complex math problems? Yes. Servers? Yes. Other things, not so much. Benchmark testing bares that much, if you go beyond geek bench scores, some applications got big bumps from more cores, others basically didn’t. Is that due to a lack of multi-core optimizations? I doubt there are many modern apps that aren’t optimized for more cores.

We are getting off track though, Apple has good CPUs now, which was my point, in ever M series Mac, and in iPhones and iPads. I just don’t see developers being CPU limited in what there is right now. Future, of course they will be. What VR/AR needs is more GPU power, not more CPU cores in most cases. Apple’s GPUs are sub-par at best, so they are going to have to focus on that. More GPU cores, more raw power, and a lot more memory. Unless the Studio or a Mac Pro brings top of the line GPU performance, we won’t be getting top of class experience with AR/VR. Maybe I’m wrong on that, depends on what resolution Apple comes with.

[chucker]

Originally Posted by PB PM Is that due to a lack of multi-core optimizations? I doubt there are many modern apps that aren’t optimized for more cores.

Well, it’s still much harder to write code that takes good advantage of multiple cores, thanks in part due to the widespread imperative approach.

[Kickaha]

Originally Posted by chucker Well, it’s still much harder to write code that takes good advantage of multiple cores, thanks in part due to the widespread imperative approach.

^--- this.

Getting back to your comment re: "IOW, it barely does" I disagree. Swift's computation model under the hood was designed for multi-core expression (not programming, expression post-compilation), LLVM was designed for runtime linking *and optimization* of computational units based on existing hardware availability, Apple Silicon is an ARM derived design that is geared for flexible shuttling of instructions and data on demand, etc, etc, etc. VE HAFF VAYZ.

This has literally been brewing for 20 years, but held back by the need to rely on Intel. Now that is dropping off rapidly, it'll be highly interesting to see what comes to fruition. Nobody right now has the groundwork laid like Apple.

PB, nobody should have to write for multiple cores, ever. The vast majority of such situations are well known, automatically detectable, and can be handled at compile time cleanly *if* the target architecture allows for clear and simple translation. That it must be currently architected and implemented that way by hand is a failure of the single core design dominance holding back the industry.

CISC winning out over RISC was a tragic misstep.

[chucker]

I usually check out when someone mentions RISC like it’s 1994.

[Frank777]

Originally Posted by Kickaha People get weird about expansion slots. :/. Somewhere they went from practical kit to a fetish.

It will be interesting to watch if the Compute Module expansion rumours turn out to be true.

On the one hand, Apple has now designed themselves into a corner TWICE with the 2013 and 2019 Mac Pros. Expansion/Upgrade modules would allow for Pro customers to upgrade systems more regularly, without throwing out the whole uber-expensive system.

On the other hand, does allowing regular internal system upgrades really sound like Apple?

[drewprops]

Originally Posted by Frank777 On the other hand, does allowing regular internal system upgrades really sound like Apple?

Maybe?

Imagine replacing your core every few years, like an iPhone...


...

[Frank777]

Maybe since a new chip sale is now revenue directly to Apple and not Intel, system upgrades will now be seen as a profit centre.

[drewprops]

Maybe it will BE your phone at some point

Insert phone A in slot B

...

[PB PM]

Originally Posted by Frank777 Maybe since a new chip sale is now revenue directly to Apple and not Intel, system upgrades will now be seen as a profit centre.

There are limits, depending on the design of the chip and the device it's going into. A new processing board could require higher peak power, which the old device couldn't deliver for example. The connection points might be different due to the amount of PCI-E lanes, other interconnects, display outputs, and so much more. For example, Intel used socket 775 for many, many years, but that didn't mean you could stick any old 775 CPU into any 775 socket motherboard and expect it to work.

[Kickaha]

Originally Posted by chucker I usually check out when someone mentions RISC like it’s 1994.

LOL Point taken. The philosophies remain though… clean design that is intended to be simple but scalable, or complexity that has diminishing returns. It’s a testament to the Intel and AMD engineers that they’ve been able to keep it going this long, but runway is running out.

[chucker]

Originally Posted by Kickaha LOL Point taken. The philosophies remain though… clean design that is intended to be simple but scalable, or complexity that has diminishing returns. It’s a testament to the Intel and AMD engineers that they’ve been able to keep it going this long, but runway is running out.

I don’t think there’s anything clean about ARM having an instruction specific to JavaScript.

And I just don’t think there’s a whole lot the compiler or CPU can do to make code scale beyond a handful of cores. You gotta take it into account as a developer.

[Frank777]

Originally Posted by Frank777 But development houses, game builders and movie studios are likely simultaneously building a host of projects for iPhone, iPad, Film/Video, Mac, and now Glasses, that would benefit from using Pro machines in their workflow. That means you need to have Pro machines to sell them.

Nice to know I'm not crazy. Apple is definitely going to debut some kind of Pro machine at WWDC.

It could be an M2 Mac Studio or early announcement of an M3 Mac Pro, but they have to put something in front of the xrOS developer market.

[Bryson]

If it isn't the long awaited Mac Pro, they might have a riot on their hands.

[chucker]

Not sure how we’re going from MacBook Air trade-in to Pro desktop.