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

691–700 of 855 posts

Re: Apple M1 Ultra

#691

Post by Tepix »

BillinghamJ wrote: post #639
ChuckMcM wrote: post #276

In some ways I wish this processor was available from a CPU chip seller. As a compute engine it gets a lot "right" (in my opinion) and would be fun to hack on. That said, the idea that USB C/Thunderbolt is the new PCIe bus has some merit. I have yet to find someone who makes a peripheral card cage that is fed by USBC/TB but there are of course standalone GPUs.

An eGPU enclosure is exactly what you're describing - PCIe x16 over TB4. They're quite commonplace now

No, TB4 is PCIe x4, unchanged since TB3.

Re: Apple M1 Ultra

#692

Post by Tepix »

toastal wrote: post #640
droopyEyelids wrote: post #298

Earlier quoted context omitted.

Friendly reminder that USB-C is a form factor, and thunderbolt is the actual transfer protocol. Sounds like you're listing the common complaints with usb-3 over usb-c peripherals, which are not a suitable replacement for PCIe. Thunderbolt is something different, more powerful & more reliable.

Thunderbolt 3 is USB4

USB4 is the successor to USB 3.2 and TB3.

Re: Apple M1 Ultra

#693

Post by GekkePrutser »

pathartl wrote: post #352
LegitShady wrote: post #299

Earlier quoted context omitted.

I always take such claims with a grain of salt anyways. It usualy on one specific benchmark. I wait for better benchmarks always instead of trusting the marketing

Even if their claims are accurate, it usually has the asterisk of *Only with Apple Metal 2. I honestly cannot understand why Apple decided they needed to write their own graphics API when the rest of the world is working hard to get away from the biggest proprietary graphics API.

Because vendor lock-in and full control over their APIs which has always been an apple staple but especially now.

5-10 years ago they were still serious about open standards, like OpenCL.. Now it's all locked in.

Re: Apple M1 Ultra

#694

Post by freemint »

2OEH8eoCRo0 wrote: post #382
gzer0 wrote: post #267

Earlier quoted context omitted.

It might also be reasonable to say that the threadripper is a workstation chip, not a chip for personal computers. Edit: even AMD themselves call their threadripper lineup workstation chips, not personal. https://www.amd.com/en/processors/workstation

I'd call them personal chips. When I think of non-personal chips I think IBM POWER or Ampere Altra.

Why do you think of Altera as non-personal chips?

Re: Apple M1 Ultra

#695

Post by Bud »

fnord123 wrote: post #653
masklinn wrote: post #632

Earlier quoted context omitted.

> assuming cooling is a solved problem I’d assume that’s what most of the chonk is about, no? > 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. The Max could be brought up to 90W or so.

I wouldn't assume heat is solved as it's been Apples weak point in the past. The cube would crack, g5 iMac's would melt capacitors, MacBooks would burn users' laps.

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.

Re: Apple M1 Ultra

#696

Post by GekkePrutser »

bradmcgo wrote: post #434
TheKarateKid wrote: post #413

Earlier quoted context omitted.

What we're seeing right now with Apple's M1 chips for desktop computing is on the same level of revolutionary as what the original iPhone did for mobile phones. The advancements in such a short period of time in the amount of computing power, low power usage, size, and heat usage of these chips is unbelievable and game-changing.

It really feels like this is all in the name of their AR/VR efforts. The killer device as far as I can think would be a simple headset that packs the capabilities of full-blown workstations. Apple Silicon seems like it could totally be on that track in some way.

True but I can only imagine the cost of that headset. It will be sure the 1% of the 1% :'(

Re: Apple M1 Ultra

#697

Post by fnord123 »

Bud wrote: post #695
fnord123 wrote: post #653

Earlier quoted context omitted.

I wouldn't assume heat is solved as it's been Apples weak point in the past. The cube would crack, g5 iMac's would melt capacitors, MacBooks would burn users' laps.

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.

Re: Apple M1 Ultra

#698

Post by hirako2000 »

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)

even their graphs don't look legit. they look like what you see in boxes of vitamins found on ransom websites. Got to wait for independent benchmarks with well explained scenarios. anything can run twice faster than anything with some specific tweaks to the program that runs.

Re: Apple M1 Ultra

#699

Post by rsynnott »

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?

The data the linker is running on will generally largely be in memory (in their posted example, which is a warm compile, completely in memory).

Re: Apple M1 Ultra

#700

Post by inkyoto »

fivea wrote: post #636
kllrnohj wrote: post #437

Earlier quoted context omitted.

To be fair that $2000 also gets you +32GB RAM, +512GB storage, and +10 CPU cores. It's not just the GPU. Although yeah you can definitely fit a 3090-equipped PC into a $4k budget even with ebay pricing if pure GPU performance is all you really want.

> To be fair that $2000 also gets you +32GB RAM, +512GB storage, and +10 CPU cores. Even though it's not an apples to apples comparison, keep in mind that a 1x32GB DIMM sells for less than 150$, and you can buy 1TB SSDs for less than 100$.

M1 Ultra is 64Gb extra, not 32.

> keep in mind that a 1x32GB DIMM sells for less than 150$

Keep in mind that M1 Pro/Max/Ultra is LPDDR5 6400 (https://www.anandtech.com/show/17024/apple-m1-max-performanc...) connected by a 512 bit memory controller.

Whereas is a kit of 2x 32 GB LPDDR5 4800 (I could not easily locate a quote for 1x 64Gb LPDDR5 4800 DIMM, leave alone 6400) retails for USD 548 (https://www.newegg.com/crucial-64gb-288-pin-ddr5-sdram/p/N82...).

I could not locate a reliable source on the type of the SSD employed in M1 Pro/Max/Ultra, so I will refrain from remarking on the comparison.

Post reply on HN