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

391–400 of 855 posts

Re: Apple M1 Ultra

#391

Post by altairprime »

bobbylarrybobby wrote: post #378
outworlder wrote: post #367

Earlier quoted context omitted.

Too bad that historically Apple has not given any attention to Mac gaming.

Part of Apple's historic MO has been to not invest in areas they don't see themselves having a competitive advantage in. Now that they can make gaming happen with a very low wattage budget they may well try to enter that space in earnest.

The difference between an Apple TV and a Mac Mini is essentially how powerful of Apple silicon it has, whether it runs tvOS or macOS, and whether it has HDMI out or not.

The Studio is a more compact form factor than any modern 4K gaming console. If they chose to ship something in that form factor with tvOS, HDMI, and an M1 Max/Ultra, it would be a very competitive console on the market — if game developers could be persuaded to implement for it.

How would it compare to the Xbox Series X and PS5? That’s a comparison I expect to see someday at WWDC, once they’re ready. And once a game is ported to Metal on any Apple silicon OS, it’s a simple exercise to port it to all the rest; macOS, tvOS, ipadOS, and (someday, presumably) vrOS.

Is today’s announcement enough to compel large developers like EA and Bungie to port their games to Metal? I don’t know. But Apple has two advantage with their hardware that Windows can’t counter: the ability to boot into a signed/sealed OS (including macOS!), load a signed/sealed app, attest this cryptographically to a server, and lock out other programs from reading with a game’s memory or display. This would end software-only online cheating in a way that PCs can’t compete with today. This would also reduce the number of GPUs necessary to support to one, Apple Metal 2, which drastically decreases the complexity of testing and deployment of game code.

I look forward to Apple deciding to play ball with gaming someday.

Re: Apple M1 Ultra

#392

Post by kllrnohj »

adfgadfgaery wrote: post #281
pphysch wrote: post #116

> This enables M1 Ultra to behave and be recognized by software as one chip, so developers don’t need to rewrite code to take advantage of its performance. There’s never been anything like it. Since when did the average developer care about how many sockets a mobo has...? Surely you still have to carefully pin processes and reason about memory access patterns if you want maximum performance.

This line is nonsense and you can safely ignore it. There have been multi-chip-modules that act like a single socket for many years. In particular, pretty much every current AMD CPU works that way. I guarantee you that for the M1 Ultra, just like every CPU before it, the abstraction will be leaky. Programmers will still care about the interconnect when eking out the last few percent of performance. Remember the Penti…

The existing AMD CPUs aren't quite like that. Technically they are all UMA, not NUMA - the L3 cache is distributed, but they are all behind a single memory controller with consistent latencies to all cores. But the Threadripper 1st gen was absolutely like that. Straight up 2+ CPUs connected via infinity fabric pretending to be a single CPU. So is that 56 core Xeon that Intel was bragging about for a while there until the 64 core Epycs & Threadrippers embarrassed the hell out of it.

Re: Apple M1 Ultra

#393

Post by jckahn »

sdenton4 wrote: post #50
iSnow wrote: post #35

Earlier quoted context omitted.

>I wonder if they'll just do 4x M1 Max for that. They'll be running out of names for that thing. M1 Ultra II would be lame, so M1 Extreme? M1 Steve?

Just need to increment the 1 instead... Eventually moving into using letters instead of numbers, until we end up with the MK-ULTRA chip.

That would really be a trip!

Re: Apple M1 Ultra

#394

Post by jedberg »

Given how low the power consumption is for the power you get, I wonder if we'll see a new push for Mac servers. In an age where reducing power consumption in the datacenter is an advantage, it seems like it would make a lot of sense.

Re: Apple M1 Ultra

#395

Post by forgotmyoldacc »

tediousdemise wrote: post #99

How does Apple manage to blow every other computing OEM out of the water? What's in the secret sauce of their company? Is it great leadership? Top tier engineering talent? Lots of money? I simply don't understand.

Apple isn't an OEM? They don't sell products that are marketed by another company.

Re: Apple M1 Ultra

#396

Post by apohn »

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…

The 3090 also can do fp16 and the M1 series only supports fp32, so the M1 series of chips basically needs more RAM for the same batch sizes. So it isn't an Oranges to Oranges comparison.

Back when that M1 MAX vs 3090 blog post was released, I ran those same tests on the M1 Pro (16GB), Google Colab Pro, and free GPUs (RTX4000, RTX5000) on the Paperspace Pro plan.

To make a long story short, I don't think buying any M1 chip make senses if your primary purpose is Deep Learning. If you are just learning or playing around with DL, Colab Pro and the M1 Max provide similar performance. But Colab Pro is ~$10/month, and upgrading any laptop to M1-Max is at least $600.

The "free" RTX5000 on Paperspace Pro (~$8 month) is much faster (especially with fp16 and XLA) than M1 Max and Colab Pro, albeit the RTX5000 isn't always available. The free RTX4000 is also a faster than M1 Max, albeit you need to use smaller batch sizes due to 8GB of VRAM.

If you assume that M1-Ultra doubles the performance of M1-Max in similar fashion to how the M1-Max seems to double the gpu performance of the M1-Pro, it still doesn't make sense from a cost perspective. If you are a serious DL practitioner, putting that money towards cloud resources or a 3090 makes a lot more sense than buying the M1-Ultra.

Re: Apple M1 Ultra

#397

Post by viktorcode »

forgotmyoldacc wrote: post #387
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]…

Neural Engine cores are not accessible for third party developers, so it'll be severely constrained for practical purposes. Currently the M1 Max is no match for even last generation mid-tier Nvidia GPU.

They are accessible to third party developers, only they have to use CoreML.

Re: Apple M1 Ultra

#398

Post by KolenCh »

pphysch wrote: post #116

> This enables M1 Ultra to behave and be recognized by software as one chip, so developers don’t need to rewrite code to take advantage of its performance. There’s never been anything like it. Since when did the average developer care about how many sockets a mobo has...? Surely you still have to carefully pin processes and reason about memory access patterns if you want maximum performance.

For any applications people use to justify buying a more than one socket machine needs this.

Eg simulation softwares often used in the industry (but the one I’ve on top of my head is Windows only.)

Anyway, the point the make is this: if you claim doubling performance, but only the selected few softwares as you observed would be optimized to take advantage of this extra performance, then this is mostly useless to the average consumer. So their point is made exactly with your observation in mind, that all your softwares is benefiting from it.

But actually their statement is obviously wrong for people in the business—this is still NUMA and your software should be NUMA aware to be really squeezing the last bit of performance. It just degrades more gracefully to non optimized code.

Re: Apple M1 Ultra

#399

Post by zitterbewegung »

jonplackett wrote: post #366
xedrac wrote: post #302

While I'm not on the Apple train, I love how they are pushing AMD, Intel and NVidia out of complacency. No more of these little tiny incremental improvements to milk the industry. Bring your BEST to the table or get left behind!

Makes you wonder how many they can glue together. Looking at you Mac Pro

Probably at least four together with LPDDR5X to get up to 1.5 TB and forty cores.
Post reply on HN