MP All Models Apple sleeping on the mac pro is comically bad timing

[Zondar]

It's time for another many-years-long "The new Mac Pro is almost certainly coming!" thread.

I think Apple never wanted to make a new Pro, but finally did after being worn down by years of wailing and gnashing of teeth (also possibly to finally bury the "trash can" memes). Oh, also because they wanted a small-scale "made in USA" product for political purposes, as I recall.

Who knows, maybe in another 5 years or whatever, as long as we all whine hard and long enough?

 

[johnalan]

you can't even buy the Studio with more than 96GB of RAM, it's not good atm for serious power users.

 

[257Loner]

Every Mac Pro since 2012 has been a "one and done" - the Trashcan, the iMac Pro, the 2019 MP (which priced out many potential buyers), the 2023 MP (same case - totally different product). I don't think Apple are particularly "strategic" about the high end).

As you pointed out, Apple has been unable to make a business case for their Mac Pro computers since 2012 whether they had Intel or Apple Silicon chips inside of them.

But you also mentioned...

What is ineffective is trying to get Apple Silicon to compete with Xeon/Threadripper/Epyc when it comes to driving multiple NVIDIA/AMD dGPUs in a big box of slots.

As you already know, Apple has not been competing against workstation CPUs and dedicated GPUs with a comparable product at a comparable price since the original Mac Pro was replaced in 2013. Yes, the 2019 Mac Pro seemed like a comparable product, but it did not sell at anywhere near a comparable price, starting at over twice the price of the original Mac Pro ($6000 vs $2500), and featuring configurations that reached $52,199.

However, it's possible that in the future the PC industry will finally catch up to Apple and obsolesce the 'big box o' slots' on the PC side as well. In 2023, Qualcomm unveiled ARM SoCs meant for laptops with their Snapdragon X series. And this past Sunday, Nvidia unveiled their ARM SoCs called RTX Spark. It remains to be seen how much of an impact on the PC industry ARM SoCs will have when competing against legacy x86 chips. But if ARM SoCs predominate the PC industry one day, then the 'big box o' slots' will have a diminished role to play on both sides of the PC/Mac divide.*

*Which is the question I believe Linus Sebastian is asking at this point (6:55) in this video:

 

[Astuces iOS]

There is as new wave of people setting up computers locally (both personal and business) for AI inference.

It is striking to me that right before this happens, Apple discontinues the Mac Pro which was perfectly suited for this. We go decades with the Mac Pro having sluggish sales and then the year that it will actually start selling like hotcakes, it's gone!

I guess Nvidia captures the market share with RTX Spark, that they announced a couple of days ago.

For the uninitiated, running a local LLM requires enough VRAM to fully fit it in memory. Models get huge very quickly. Because of the unified memory, this means you can fit much larger models on a mac instead of a traditional machine that has a separate video card (the video card has its own RAM and video cards with a lot of RAM are very, very expensive)

Simple question, how do you make the Mac Pro carbon neutral?

 

[mattspace]

This seems completely unrealistic with all that apple's been saying and doing the last decades. The iPad is closed so the Mac can stay open.

that's an exceptionally naïve take on the situation; the sort of thing Gruber would say after Apple rented his mouth.

The Mac is just a UI skin for an iPad; what Apple has been doing with it, is locking down consistently to be more iPad-like, and less Mac-like.

 

[mattspace]

Simple question, how do you make the Mac Pro carbon neutral?

Use it for a decade on extra commissioned solar power to offset the embodied energy of manufacturing the product (and the solar infrastructure) in the first place.

Nothing that becomes obsolescent in sub-decadal timescales is carbon neutral.

 

[Astuces iOS]

Use it for a decade on extra commissioned solar power to offset the embodied energy of manufacturing the product (and the solar infrastructure) in the first place.

Nothing that becomes obsolescent in sub-decadal timescales is carbon neutral.

Still, I don't think it will follow what apple said for 2030

 

[ignatius345]

The Mac is just a UI skin for an iPad

That’s... demonstrably false, but whatever.

 

[mattspace]

Still, I don't think it will follow what apple said for 2030

"blah blah licking the boot of the uniformed man in the Hugo Boss suit"?

 

[mattspace]

That’s... demonstrably false, but whatever.

aside from the part where they replaced the macOS mass-storage infrastructure with the one from iOS / iPadOS and broke PCI NVME support for the entirety of the Sonoma release, because iPads don't have to take into account external storage at pre-boot... but sure you keep telling yourself that.

 

[ignatius345]

People love to get all up in arms about which Macs they think Apple should be building, but you’re talking about a 7.6% segment of their revenue. And the vast majority of that 7.6% is from MacBooks. That leaves, what, maybe 2% of their overall revenue coming from these desktops they’re doomed because they didn’t bring to market? Please.

 

[splifingate]

What is ineffective is trying to get Apple Silicon to compete with Xeon/Threadripper/Epyc when it comes to driving multiple NVIDIA/AMD dGPUs in a big box of slots.

Aye; particularly highlighted by the fact that we can SSH/HTTP in to these big boxes remotely on devices that just need enough CPU to fuel a browser and a terminal . . . heck, I can even tap-in to my personal (local) LLMs with my phone on a mountain in Wyoming if I want 🙂

My Big Box has AMD dGPUs right now, and can be switched-out with others (in the future).

All well-and-good, but I'd rather have a Big Unified Mx 😉

 

[HuskerNKS]

Simple question, how do you make the Mac Pro carbon neutral?

By putting 8 GB of RAM in it. You ok with that?

 

[leman]

I don’t understand what OP is trying to say. The Mac Pro uses the same hardware and offers identical storage capacities as the Studio. Apple already offers a device similar to RTX Spark, and has done so for years.

 

[RedWeasel]

I guess Nvidia captures the market share with RTX Spark, that they announced a couple of days ago.

I could very well imagine that happening. RTX Spark has two huge advantages: CUDA and Windows. The latter might be controversial, but in an enterprise setting, this is absolutely a plus.

Those Mac Studio setups are nice for proof of concepts or personal projects or maybe small businesses, but larger companies will appreciate all the controls and management you get with Windows.

 

[Bug-Creator]

I don’t understand what OP is trying to say. The Mac Pro uses the same hardware and offers identical storage capacities as the Studio. Apple already offers a device similar to RTX Spark, and has done so for years.

I see 2 way a MPro could make "sense" for local F(ake)I.

- add a bunch of NVidia GPUs do the work, can (could/should?) be done with the M2-Ultra version, and could be done with putting them in a Thunderbolt case connected to a MacStudio

- create a custom M6-HypaHypa with 4 M6-Max dies or even an M6-Über with 8 and add 1-2 TB of RAM. Even without the RAM shortage the lack of economics of scale for that chip is simple making it a non starter.

In short just a usual case of someone getting capture by a hype machine failing to understand specs or how tech works in general.

 

[RedWeasel]

Times change. When the XServe launched, it was competing against Windows Server, Netware and commercial Unix with sky-high (often "per seat") licensing - and Macs were reliant on various proprietary Apple protocols. The market was killed by Linux, turnkey NAS products with friendly web-based config screens and a move towards open networking standards (including the opening-up of SMB/CIFS).

Funnily enough, early Mac OS X Server was initally using Samba for SMB and only switched to Apple's own SMBX implementation due to the change of Samba to GPLv3.

 

[RedWeasel]

Meanwhile, Apple Silicon has proven to be a major strategic advantage at the low/mid-range all the way up to the Studio. I think the "high end" is something they're happy to sacrifice - they don't need to have a foot in every pie.

I think Apple also wouldn't have the software stack to compete in the high end. At least not anymore.

 

[leman]

I could very well imagine that happening. RTX Spark has two huge advantages: CUDA and Windows. The latter might be controversial, but in an enterprise setting, this is absolutely a plus.

Those Mac Studio setups are nice for proof of concepts or personal projects or maybe small businesses, but larger companies will appreciate all the controls and management you get with Windows.

For ML work? Why would you use Windows if most of tooling and infra is Unix/linux-first? While Windows works for classical corporate environments, I fail to see any benefit for ML development.

 

[leman]

I think Apple also wouldn't have the software stack to compete in the high end. At least not anymore.

What does “high end” mean in this context? Training and deploying huge LLMs? Definitely out of scope for them (although they do have custom ML servers for running large models).

 

[Dirk Pieters]

People love to get all up in arms about which Macs they think Apple should be building, but you’re talking about a 7.6% segment of their revenue. And the vast majority of that 7.6% is from MacBooks. That leaves, what, maybe 2% of their overall revenue coming from these desktops they’re doomed because they didn’t bring to market? Please.

View attachment 2634855

I was thinking the same thing. I remember the Pro had quite a lot of options during config. It's more complex machine to build than a Studio for example. I bet the Pro's sales numbers have been dwindling for years and at some point it just isn't worth the effort anymore.

If the market has decided with their wallets that they don't want Pro machines anymore for whatever reason, then why keep it alive? Apple is relentless in ditching old tech.

 

[RedWeasel]

For ML work? Why would you use Windows if most of tooling and infra is Unix/linux-first? While Windows works for classical corporate environments, I fail to see any benefit for ML development.

Well the tooling is mostly Python and C++ or rather CUDA based and pretty agnostic of the OS beneath it. And Microsoft has many interesting projects like MXC for example (in addition to WSL 2).

So one advantage that Windows has (IMHO) is that you can run all these tools within a corporate environment, i.e. with all the controls, authentication etc. that comes with those environments.

Also, a lot of knowledge in big corporations is in Microsoft systems.

So for inference, Windows absolutely is an interesting platform I think (at least for businesses).

 

[SDAVE]

There is as new wave of people setting up computers locally (both personal and business) for AI inference.

It is striking to me that right before this happens, Apple discontinues the Mac Pro which was perfectly suited for this. We go decades with the Mac Pro having sluggish sales and then the year that it will actually start selling like hotcakes, it's gone!

I guess Nvidia captures the market share with RTX Spark, that they announced a couple of days ago.

For the uninitiated, running a local LLM requires enough VRAM to fully fit it in memory. Models get huge very quickly. Because of the unified memory, this means you can fit much larger models on a mac instead of a traditional machine that has a separate video card (the video card has its own RAM and video cards with a lot of RAM are very, very expensive)

I don't think Apple cares about this segment at a "Server" level. That's why they allow daisy chaining via Thunderbolt multiple Mac Studios for example.

NVIDIA is producing enterprise level AI infrastructures and workflows with very propietary layers like CUDA. You have to be in their ecosystem. Apple isn't doing any of that, it just happens to be that Apple Silicon can run LLMs.

At the end of the day, Apple is a consumer product company.

Also to be honest with you, NVIDIA may be leading things right now, but lower cost chips/less power hungry chips are going to go towards TPUs vs GPUs. Google literally does not use NVIDIA, the rest will follow.

NVIDIAs stock is propped up by a bubble, and Apple doesn't dive into an industry as unstable as that.

 

[Contact_Feanor]

Use it for a decade on extra commissioned solar power to offset the embodied energy of manufacturing the product (and the solar infrastructure) in the first place.

Nothing that becomes obsolescent in sub-decadal timescales is carbon neutral.

No amount of solar energy use “offsets” carbon emissions. That’s not how it works.

 

[mattspace]

No amount of solar energy use “offsets” carbon emissions. That’s not how it works.

If work that was going to get done anyway with fossil fuels is done with a solar system, that is an offset. If you avoid the embodied energy costs of causing a new thing to be manufactured, by keeping an old thing in service, that is an offset.