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

461–470 of 551 posts

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

#461

Post by alpaca128 »

nox101 wrote: post #347
alpaca128 wrote: post #310

Earlier quoted context omitted.

Yes, but you only really encounter that when pushing the CPU to 100% for more than a few minutes. The cooling is objectively terrible, but still easily enough for most users, that's the crazy thing.

maybe? as local LLM/SD etc get more common it might be common to push it. I've been getting my fans to come on and get burning hot quite often lately because of new tech. I get that I'm a geek but with Apple, Google and everyone else trying to run local ML it's only a matter of time.

Apple's chips already have AI accelerators for things like content-based image search. They would never retroactively worsen battery life and performance just for a few more AI features when they could instead use it as selling point for the next hardware generation.

And if you regularly use local generative AI models the Pro model is the more reasonable choice. At that point you can forget battery life either way.

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

#462

Post by nabakin »

whynotminot wrote: post #141
nabakin wrote: post #134

Earlier quoted context omitted.

In power efficiency maybe, but not top performance

Literally yes top single core performance. (And incidentally also efficiency)

I stand corrected, that's incredible.

https://www.cpubenchmark.net/singleThread.html

https://www.cpu-monkey.com/en/cpu_benchmark-cinebench_r23_si...

https://www.cpubenchmark.net/high_end_cpus.html

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

#463

Post by PoignardAzur »

0x457 wrote: post #299
robotresearcher wrote: post #288

Earlier quoted context omitted.

Are you saying it's faster to build a binary with native architecture? Why is that?

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?

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

#464

Post by deafpolygon »

wing-_-nuts wrote: post #44
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.

I do wonder if PC desktops will eventually move to a similar design. I have a 7800x3d on my desktop, and the thing is a beast but between it and the 3090 I basically have a space heater in my room

After having my PC for (almost) 4 years, I can say that this beast is the last large form computer I will buy.

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

#465

Post by jocaal »

m00x wrote: post #371
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?

Did no one check the scores? They're not the top consumer CPU by quite a range. It's probably the best power per watt, but not the most powerful CPU.

The performance-per-watt isn’t necessarily the best. These scores are achieved when boosted and allowed to draw significantly more power. Apple CPUs may seem efficient because, most of the time, computers don’t require peak performance. Modern ARM microarchitectures have been optimized for standby and light usage, largely due to their extensive use in mobile devices. Some of MediaTek and Qualcomm's CPUs can offer better performance-per-watt, especially at lower than peak performance. The issue with these benchmarks is that they overlook these nuances in favor of a single number. Even worse, people just accept these numbers without thinking about what they mean.

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

#466

Post by ddingus »

MBCook wrote: post #378
ddingus wrote: post #185

Earlier quoted context omitted.

My 2010 i7 MBP would do that under heavy loads. All aluminum body, with fans, and when that CPU really had to work it put out a lot of heat. Compiling gcc multi thread a few times would be enough.

Not only that, they seemed to get hotter with each subsequent generation. My 2015 could get hotter than my 2010. I think my work 2019 gets hotter still. I think the Intels were hotter than my G4, but it’s been too long and the performance jump was worth it. Got an M1 Air, it blows them all out of the water (even 2019, others aren’t a surprise). And it does it fanless, as opposed to emulating a jet engine. And if you…

My experience is the same. I only owned one Intel MacBook Pro.

Was the only one I needed thankfully. Missed the whole port reduction and power bar mess.

I love my M1 Air. It is the first general purpose computing hardware that felt like a real advance. I measured that two ways:

How much closer to my mobile is it?

How much faster is it?

The Air feels like a Mobile Computer if that makes any sense. One USB port expander to serve as a dock of sorts later and it makes for a great desktop experience.

When using it on the go, it has that light, powerful feel much like running my phone does.

Great machine. It is easily my favorite computer Apple has ever made, 8 bit greatness and an older age aside.

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

#467

Post by wslh »

mortenjorck wrote: post #27
post_break wrote: post #8

What's crazy is that the M4 Pro is in the Mac mini, something so tiny can handle that chip. The Mac Studio with the M4 Max will be awesome but the Mini is remarkable.

Indeed, the $1599 M4 Pro Mac mini beats the current $3999 M2 Ultra Mac Studio on GeekBench 6 multi-core: https://www.macrumors.com/2024/10/31/m4-pro-chip-benchmark-r...

Personally, the Mac Mini will be my reentry into desktop computers after more than 1.5 decades [1]. Your comment got me thinking: could this be another perfectly calculated move by Apple? After all, I’ve only bought mobile devices from them until now. I’m eager to see Apple’s financial results for Q4 2024 and Q1 2025 to understand how this strategy plays out.

[1] https://news.ycombinator.com/item?id=41996008

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

#468

Post by spockz »

MBCook wrote: post #392
spockz wrote: post #254

Earlier quoted context omitted.

For many workloads I think they are pulling definitely ahead. However, I think there is still much to gain in software. For example, my Linux/Fedora desktop with 5900X is many times more responsive than my 16” M1 Pro. Java runs faster. GraalVM native generated native images run way waster. Golang runs faster. X86_64 has seen more love from optimalisations than aarch64 has. One of the things I hit was different GC/mem…

I suspect that’s an OS issue. Linux is simply more optimized and faster at equivalent OS stuff. Which isn’t too surprising given a lot of the biggest companies in the world have been optimizing the hell out of it for their servers for the last 25+ years. On the flipside of the coin though Apple also clearly optimizes their OS for power efficiency. Which is likely paying some good dividends.

For sure, the networking is on the OS level because in both I’m using the tcp stack provided by the OS. (Also GUI in Linux has less transitions and feels snappier in general.)

The remainder can be attributed to compiler optimisations or lack thereof.

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

#469

Post by mkesper »

TacticalCoder wrote: post #360
philistine wrote: post #120

Earlier quoted context omitted.

Back in the early 90s, when Apple was literally building their first product line with the Mac, they would come out with their second big honking powerhouse Mac: the Macintosh IIx. It blew everything out of the water. Then they would come out with their next budget all-in-one machine. But computing was improving so fast, with prices for components dropping so quickly, that the Macintosh SE/30 ended up as impressive a…

> I thought the M1 Macbook Air I've got one and it's really not that impressive. I use it as a "desktop" though and not as a laptop (as in: it's on my desk hooked to a monitor, never on my laps). I'm probably gonna replace it with a Mini with that M4 chip anyway but... My AMD 7700X running Linux is simply a much better machine/OS than that MacBook M1 Air. I don't know if it's the RAM on the 7700X or the WD-SN850X SSD…

There are tons of affordable and upgradable NUC form factor machines for running Linux, so why a Mac Mini (if you're not running LLMs locally, then the good support and fast integrated RAM might be a reason)?

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

#470

Post by rbanffy »

MBCook wrote: post #391
rbanffy wrote: post #226

Earlier quoted context omitted.

> Edit: I am looking at peak 1T performance, not efficiency. In that regard I don't think anyone has been close. Indeed. Anything that approaches Apple performance does so at a much higher power consumption. Which is no biggie for a large-ish desktop (I often recommend getting middle-of-the-road tower servers for workstations).

Don’t thermals basically explode non-linearly with speed? It’s possible Apple’s chips could be dramatically faster if they were willing to use up 300W. I remember seeing an anecdote where Johny Srouji, the chief Apple Silicon designer, said something like the efficiency cores get 90% of the performance of the performance cores at like 10% of the power. I don’t remember the exact numbers but it was staggering. While t…

I’m now imagining an “M4 Hot” for the next MacPro with an Intel Xeon W power budget.

> they could (theoretically) make a chip of only 32 efficiency cores and just sip power.

Intel and AMD did that with their latest data-center chips to compete with Ampere and AWS’s Graviton. I’d love to build a workstation out of one such beast.

Post reply on HN