Live data from Hacker News

Apple M1 Ultra

apple.com

121–130 of 855 posts

Re: Apple M1 Ultra

#121

Desktop marketing seems to be getting desperate. Few people use the apps they show in the Mac Studio benchmarks. Fewer still care that their chips use less power... if they did, they would stay on their phones.

Large businesses do. I bet in some places you could get grant/loan incentives to replace a PC fleet with these things. Back in the Pentium-4 days, iirc I was able to get almost $250k in grants and $1.5M in subsidized loans to do accelerated refresh of a PC fleet and small datacenter, all through a utility's peak load reduction program.

I don't deny such things happen but it's illogical. If you have 20 people on desktops a small fraction of them will use more energy for microwaving lunch, making coffee or on air conditioning than they will save in aggregate on this nominal reduction in power draw.

Re: Apple M1 Ultra

#122

I read “Apple unveils MK ULTRA”

Low key what if this was planned to change Google results to "Did you mean M1 Ultra?" when searching for the experiment? The CIA is using all that money for something consumers can use now!

/takes off foil hat

Re: Apple M1 Ultra

#123
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.

Right. Still riding gains from the node shrink and on package memory. Could AMD/Intel follow suit and package memory as an additional layer of cache? I worry that we are being dazzled by the performance at the cost of more integration and less freedom.

The next CPU coming from AMD will be the 5800X3D with 96MB cache. They stack 64MB L3 on top. Rumours say it comes out 20th of April.

edit: typo + stacking + rumoured date

Re: Apple M1 Ultra

#124
post #95

Earlier quoted context omitted.

Based on Anandtech benchmarks the M1 Max GPU is basically on par with a mobile 3080, which a quick search tells me is about 60% as fast as a desktop 3080. Not unreasonable to believe 2 of them combined will outperform a 3090- with nearly 128 GB of VRAM to boot. Even more incredible Anandtech reports the M1 max GPU block maxing at 43W in their testing. So a 90W GPU in the M1 Ultra is trading blows with a 350+ watt 309…

GPU scaling is absolutely not linear in that way. nvidia gave up on that in recent generations as without software support to match, you had situations where double 1080s were 95% as fast as one 1080 with worse frame times. Might be nice for e.g. ML where you can effectively treat them as entirely independent GPUs but for games I would be surprised if this matches a high end GPU.

macOS will see it as one gpu.

Re: Apple M1 Ultra

#126
post #69
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?

I'm pretty sure all their messaging is preparing us for "M1 Outrageous"

M1 Ludicrous the IV

Re: Apple M1 Ultra

#127

Didn't AMD do something similar with putting 2 CPU chips together with cache in-between? What's the difference here in packaging technology? (maybe there is no shared cache here)

Yes. AMD has had integrated CPU+GPU+cache-coherent HBM for a while. You can't buy these parts as a consumer though. And they're probably priced north of 20k$/each at volume, with the usual healthy enterprise-quality margins.

Re: Apple M1 Ultra

#128

Earlier quoted context omitted.

How is slotted RAM slower? 6400Mhz DIMM exists. This would match the specs of the RAM on the M1 Max. Even octa-channel has been done before so the memory bus would have the exact same width, latency and clock frequency.

The memory bandwidth of the M1 Max is 400 GB/s with 64GB of RAM, where as the memory bandwidth of Corsair's 6400MHz DDR5 32GB RAM module is 51GB/s per stick, or 102GB/s for the M1 Max equivalent.

51GB/s * 8 (octa-channel, not dual channel as you are calculating) is 408 GB/s. Basically the same as the M1 Max. It's not fair to use an off the shelf product since even if the RAM is slotted Apple wouldn't use an off the shelf product.

Whether they use slotted RAM or not has nothing to do with performance. It's a design choice. For the mobile processors it makes total sense to save space. But for the Mac pro they might as well use slotted RAM. Unless they go for HBM which does offer superior performance.

Re: Apple M1 Ultra

#129
post #65

Earlier quoted context omitted.

The Mac Pro will have replaceable RAM. It will use the RAM soldered onto the CPU as cache. You’ll most likely also be able to buy dedicated GPUs/ML booster addon Cards and the likes for it. It’s the most likely thing to happen or they won’t release another Mac Pro.

Why would they use soldered RAM as cache? It's not like it's faster then replaceable RAM. Unless they go HBM2 but I doubt that.

The bandwidth of the soldered ram is much higher which makes it much faster for code that accesses a lot of RAM like video editors.

Re: Apple M1 Ultra

#130

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.

I think it's mostly engineering and the cash to make things happen. You heard it today in the presentation that since they launched M1, sales have skyrocketed for Apple computers.

Hopefully leadership is really looking hard at this trend and adjusting future offerings accordingly. Consumers WANT machines with high performance and great I/O and they're willing to pay for them.

With Apple, Intel, and AMD really stepping up the last couple of years, I think the next decade of personal computing is going to be really exciting!

Post reply on HN