Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple M1 Ultra

apple.com

Post reply on HN
Print view View original

711–720 of 855 posts

Re: Apple M1 Ultra

#711

Post by threeseed »

anaisbetts wrote: post #678
kllrnohj wrote: post #360

Earlier quoted context omitted.

The GPU claims on the M1 Pro & Max were, let's say, cherry picked to put it nicely. The M1 Ultra claims already look suspicious since the GPU graph tops out at ~120W & the CPU graph tops out at ~60W yet the M1 Studio is rated for 370W continuous power draw. Since you mention ML specifically, looking at some benchmarks out there (like https://tlkh.dev/benchmarking-the-apple-m1-max#heading-gpu & https://wandb.ai/tcapel…

At some point can we finally admit that Apple's GPU claims just aren't....true? Like, every Apple keynote they put up incredible performance claims, and every time people actually get their hands on the product, it doesn't even come close to holding water in any domain where GPU performance matters (Game performance, ML training perf)

> GPU performance matters (Game performance, ML training perf)

No one plays games on a Mac.

And it has nothing to do with GPU performance but rather the fact that the audience simply isn't interested in gaming on it and so there is no economic incentive to target them.

So the GPU performance that matters to Mac users and is relevant to Apple is not games but rather content creation, production etc.

Re: Apple M1 Ultra

#712

Post by nicoburns »

fivea wrote: post #631
nicoburns wrote: post #295

Earlier quoted context omitted.

https://github.com/rui314/mold would suggest otherwise. Massive speedups by multithreading the linker. I think traditional linkers just aren't highly optimised.

> https://github.com/rui314/mold would suggest otherwise. Does it, though? I mean, if you read that link you'll notice it boasts the linker's performance by comparing it with cp and how it's "so fast that it is only 2x slower than cp on the same machine." Is cp supposed to be CPU-bound?

Yes, but that's for mold, which is multithreaded. The original context of this thread being the question of whether a linker would see speedups from multithreading. Most people are using traditional single-threaded linkers which are an order of magnitude slower than mold. The fact that mold is so much faster suggests that a linker does indeed see big speedups from multithreading.

Re: Apple M1 Ultra

#713

Post by jamesfmilne »

ChuckMcM wrote: post #326
jamesfmilne wrote: post #292

Earlier quoted context omitted.

https://www.sonnettech.com/product/thunderbolt/pcie-card-exp...

Thanks! Of course they are bit GPU centric but the idea is there. Very interesting stuff. I wonder both if the Zynq Ultrascale RFSOC PCIe card would work in that chassis and if I could get register level access out of MacOS.

Yes, you can interface with PCIe devices using DriverKit, Apple's new user-space device driver platform.

No need to run inside the kernel for these things any more.

https://developer.apple.com/documentation/driverkit

https://developer.apple.com/documentation/pcidriverkit

Re: Apple M1 Ultra

#714

Post by jamesfmilne »

Synaesthesia wrote: post #96

Apple have some serious chip design abilities. Imagine if they entered the server market, with this architecture it could be very successful.

My guess is they will launch their own cloud platform powered by Apple Silicon.

Click & deploy from within Xcode (I hate Xcode though.)

Re: Apple M1 Ultra

#715

Post by truculent »

cube2222 wrote: post #4

Looks like all the people saying "just start fusing those M1 CPU's into bigger ones" were right, that's basically what they did here (fused two M1 Max'es together). And since the presenter mentioned the Mac Pro would come on another day, I wonder if they'll just do 4x M1 Max for that.

My guess would be an M1 Ultra with additional modular expansions for specific purposes. I.e. instead of having a GPU, you could add a tensor processing module if you need to do machine learning, or a graphics processing chip for video production, and so on.

Re: Apple M1 Ultra

#716

Post by geraldwhen »

kllrnohj wrote: post #360
polyrand wrote: post #258

I think the GPU claims are interesting. According to the graph's footer, the M1 Ultra was compared to an RTX 3090. If the performance/wattage claims are correct, I'm wondering if the Mac Studio could become an "affordable" personal machine learning workstation (which also won't make the electricity bill skyrocket). If Pytorch becomes stable and easy to use on Apple Silicon [0][1], it could be an appealing choice. [0]…

The GPU claims on the M1 Pro & Max were, let's say, cherry picked to put it nicely. The M1 Ultra claims already look suspicious since the GPU graph tops out at ~120W & the CPU graph tops out at ~60W yet the M1 Studio is rated for 370W continuous power draw. Since you mention ML specifically, looking at some benchmarks out there (like https://tlkh.dev/benchmarking-the-apple-m1-max#heading-gpu & https://wandb.ai/tcapel…

Gigabyte and strix 3090 are routinely in stock at Newegg. Msrp. The shortage is over

Re: Apple M1 Ultra

#717

Post by hughrr »

fnord123 wrote: post #697
Bud wrote: post #695

Earlier quoted context omitted.

Those issues were all almost 20 years ago. If you think this is still a problem, you haven't used any recent Macs. The current MB Air and MB Pro both run very cool even under prolonged heavy loads. Apple's management of any and all heat issues has been far better than any competitors for a while now.

The last intel macs run way too hot still. They get up over 80dC. The 2014 ones would hard reboot when building large Java projects.

To be fair the current Dells do the same.

Re: Apple M1 Ultra

#718

Post by kllrnohj »

threeseed wrote: post #711
anaisbetts wrote: post #678

Earlier quoted context omitted.

At some point can we finally admit that Apple's GPU claims just aren't....true? Like, every Apple keynote they put up incredible performance claims, and every time people actually get their hands on the product, it doesn't even come close to holding water in any domain where GPU performance matters (Game performance, ML training perf)

> GPU performance matters (Game performance, ML training perf) No one plays games on a Mac. And it has nothing to do with GPU performance but rather the fact that the audience simply isn't interested in gaming on it and so there is no economic incentive to target them. So the GPU performance that matters to Mac users and is relevant to Apple is not games but rather content creation, production etc.

M1's blender rendering performance also wasn't very good, though. Assuming the scene was small enough to fit in the smaller VRAM of competitive consumer GPUs anyway.

Video content creation specifically is where it mostly achieves what the graphs indicate, and that's mostly a "well yeah, video decoder ASICs are really efficient. I'll take things we already knew for $100, Alex"

Re: Apple M1 Ultra

#719

Post by ChrisRR »

Geee wrote: post #307

This is insane. They claim that its GPU performance tops the RTX 3090, while using 200W less power. I happen to have this GPU on my PC and not only it costs over $3000, but its also very power-hungry and loud. Currently, you need this kind of GPU performance for high resolution VR gaming at 90 fps, but its just barely enough. This means that the GPU will run very loudly and heat up the room, and running games like HL…

You're right, it is insane, but this is Apple.

So expect what that graph actually means is some extremely specific, cherry-picked benchmarks

Re: Apple M1 Ultra

#720

Post by kllrnohj »

dingle_thunk wrote: post #623
brigade wrote: post #385

Earlier quoted context omitted.

100W of that is probably for the USB ports; afaik TB4 ports are required to support 15W, and I don’t think there’s been a Mac didn’t support full power simultaneously across all ports. (if that’s even allowed?)

I suppose given that this is two M1 Max glued together, assuming cooling is a solved problem, the max SOC power consumption just twice as high as usual, plus interconnect overhead. Right? Based on the thermal and power consumption characteristics of previous chips I would not be surprised if say ~120W is the max power draw of this thing. edit: Of course the M1 max only shipped in laptops, so... who knows.

The M1 Max hits 100w in a laptop form factor with 'real' workloads when hitting CPU and GPU simultaneously (or at least not parasitic ones like prime95 & furmark). So this is probably >200w, unless it's been power limited and thus performs worse than 2x M1 Max's do anyway.
Post reply on HN