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

Apple's M4 Max chip is the fastest single-core performer in consumer computing

twitter.com

Post reply on HN
Print view View original

471–480 of 551 posts

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#471

Post by kapoolo »

PoignardAzur wrote: post #463
0x457 wrote: post #299

Earlier quoted context omitted.

Because there two ways to get to the same result: - use native toolchain to produce artifacts for a different architecture - use emulation to run different arch toolchain to produce different arch artifacts First one is fast, second one is slow. In docker only second variant is possible.

That doesn't sound right. It's not like you need to run in WebAssembly to produce WebAssembly binary. Why would you need to emulate x86 to produce x86-shaped bytes?

Emulating the target architecture's SDK can be much easier and simpler to setup and avoid mistakes. You do not need to make any changes or configuration to make it compile.

For exmaple, that is generally the way you cross compile Flatpaks in cli or ide. In, for example, GNOME Builder you can just select the device you want to build for, like your smartphone, and it uses QEMU to emulate the entire SDK in the target architecture, you can also seemlessly run the Flatpak on your host through QEMU user mode emulation too.

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#472

Post by rbanffy »

TinkersW wrote: post #264
rbanffy wrote: post #221

Earlier quoted context omitted.

> If there are a lot of game simulation calculations Why not move at least some of that into the GPU as well? Lots of different branchy code paths for the in-game objects?

Latency of communication with GPU is too large, and GPUs suck at things CPUs are good at.

Wouldn’t the latency be because of no shared memory and game context data having to go over the PCIe bus?

GPUs suck at branchy code, but is that the case with this simulation data? I can see why it’d suck for NPC simulation, but not physics.

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#474

Post by janwas »

BoingBoomTschak wrote: post #230
llm_nerd wrote: post #152

Earlier quoted context omitted.

Geekbench is an excellent benchmark, and has a pretty good correlation with the performance people see in the real world where there aren't other limitations like storage speed. There is a sort of whack-a-mole thing where adherents of particular makers or even instruction sets dismiss evidence that benefits their alternatives, and you find that at the root of almost all of the "my choice doesn't win in a given benchm…

Any proprietary benchmark that's compiled with the mystery meat equivalent of compiler/flags isn't "excellent" in any way. SPECint compiled with either the vendor compiler (ICC, AOCC) or the latest gcc/clang would be a good neutral standard, though I'd also want to compare SIMD units more closely with x265 and Highway based stuff (vips, libjxl). And how do you handle the fact that you can't really (yet) use the same…

Good points. gemma.cpp can also be an interesting benchmark, it also uses Highway.

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#475

Post by jchw »

mabedan wrote: post #440
jchw wrote: post #366

Well, firstly, it isn't. There are higher Geekbench 6 CPU scores, even forgetting the ones that appear to be fake or at least broken. But secondly, that would absolutely not indicate that it is the "fastest single-core performer in consumer computing". That would indicate that it is the highest scoring Geekbench 6 CPU in consumer computing. Whether or not that's actually a good proxy for the former statement is up to…

> you really need to be able to compare CPUs is a healthy mix of synthetic benchmarks and real-world workloads I like how Apple got roasted on every forum for using real world workloads to compare M series processors to other processor. The moment there’s a statistic pointing to “theoretical” numbers, we’re back to using real world workload comparison

You need both, this is just standard practice. Check for example how CPUs are compared on outlets like Gamer's Nexus or Phoronix.

Apple didn't get roasted for presenting real world performance, they got roasted for doing the kinds of things that marketing people do: making vague blanket claims about performance that couldn't actually be reasonably validated. (Intel and AMD routinely get roasted for similar things.)

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#476

Post by kristianp »

varispeed wrote: post #208
carlgreene wrote: post #11

Earlier quoted context omitted.

I am still astounded the huge change moving from an Intel Mac to an Apple Silicon Mac (M1) has had in terms of battery performance and heat. I don't think i've heard the fans a single time I've had this machine and it's been several years. Nor have I had any desire to upgrade.

On extremely heavy workloads the fans do engage on my M1 Max, but I need to get my ear close to the machine to hear them. Recently my friend bought a laptop with Intel Ultra 9 185h. It roared fans even when opening Word. That was extraordinary and if it was me making the purchase I would have sent it back straight away. My friend did fiddle a lot with settings, had to update BIOS and eventually the fan situation was…

I have an i7 12th gen thinkpad and the fans would often be audible when I first got it with Windows 11. Then I installed linux. Now the fan is only audible when something has gone wrong and a process (usually chrome) is pinning a core.

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#477

Post by pxmpxm »

abhinavk wrote: post #247
pxmpxm wrote: post #231

Earlier quoted context omitted.

Here's the geekbench link https://browser.geekbench.com/v6/cpu/8593555 How/where are they getting 128gb of ram? I don't see that as an option for any of the pre-orders pages. Still pretty impressive, I get 1217/10097 with dual xeon gold 6136 that doubles as a space heater in the winter.

Switch to M4 Max 16-core CPU, it will unlock 64 and 128GB options for memory.

Mac minis aren't available with set up - seems only macbook pro?

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#478

Post by Damogran6 »

eastbound wrote: post #197
spiderfarmer wrote: post #82

Earlier quoted context omitted.

My best Apple purchases in 20 years of being their customer: The Macbook M1 Pro 16 inch and the Pro Display XDR. When Steve Jobs died I really thought Apple was done, but their most flawless products (imho) came much later.

Yeah, don’t forget the 10 dark years between the Butterfly Keyboard Macbook Pro 2016, the Emoji Macbook Air, until the restoration of the USB ports… around 2022. That was truely the dark age of Apple.

That's the common hipster take on it...but I kinda liked the way the butterfly keys felt and my impression of the touchbar was chaotic neutral. By the time they let up a little on the ports, I'd lived with 4 USB-C ports so long that it really wasn't that big a deal.

What got me, however, was that was the time where their trade-in program was really kicking in. I think I got $800 for my touchbar mac which made the jump to an M1 Pro 14 a little less painful. Now you don't seem to so much pay for hardware as lease the Apple experience, so long as the hardware is still good.

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#479

Post by kalleboo »

TylerE wrote: post #447
kalleboo wrote: post #446

Earlier quoted context omitted.

Yeah I remember that, I posted a YouTube video complaining about it 6 years ago, before I could find any other references to the issue online. https://www.youtube.com/watch?v=Rox2IfViJLg That would cause it to throttle even when idle! But even on battery or using the right-hand ports, under continuous load (edit-build-test cycles) it would quickly throttle.

Oh yeah, I'm very aware. My work machine was a 2015 MBP until about 6 months ago. It was really bad towards the end.

Congratulations on upgrading away from it!

Re: Apple's M4 Max chip is the fastest single-core performer in consumer computing

#480

Post by quitit »

eigenspace wrote: post #4

I'll admit to some reflexive skepticism here. I know GeekBench at least used to be considered an entirely unserious indicator of performance and any discussion relating to its scores used to be drowned out by people explaining why it was so bad. Do those criticisms still hold? Are serious people nowadays taking Geekbench to be a reasonably okay (though obviously imperfect) performance metric?

Benchmarking itself is of limited usefulness, since in reality it is only an internally-relevant comparison.

Translating these scores into the real world is problematic. There are numerous examples of smart phones powered by Apple chips versus Qualcomm chips having starkly different performance with actual use. This is in spite of the chips themselves scoring similarly in benchmarks.

The interesting thing here isn't really how high it's scored against other chip brands, but how it outperformed the M2 Ultra. There was some hum of expectation on HN that the differences between M1, M2, M3 etc would be token and that Apple's chips devision is losing its touch. Yet the M2 Ultra in the Mac Studio was released in June 2023, and the M4 Pro in the mini now for November 2024. That is quite the jump in performance over time and a huge change in bang for buck.

Post reply on HN