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

16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

twitter.com

Post reply on HN
Print view View original

511–520 of 529 posts

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#511

Post by tambourine_man »

utxaa wrote: post #504
tambourine_man wrote: post #71

Earlier quoted context omitted.

I lot of people are having a hard time grasping or accepting. “But this benchmark has these issues”, “but this is hardware accelerated, so it's not a fair comparison”, etc. I think when enough benchmarks and real world usage have been run, it will sink in. What did people expect? They've been killing it with 5W fan-less chips for years. Have you seen how confident Apple is in those videos?

it's fair to be careful with something new. if it's that great and it will be quickly apparent.

I wouldn't classify the backlash as carefulness.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#512

Post by ddingus »

DonaldPShimoda wrote: post #507
ddingus wrote: post #243

Earlier quoted context omitted.

Seems to me, off chip RAM becomes a new sort of cache. If Apple sizes the on chip RAM large enough for most tasks to fly, bigger system RAM can get paged in and performance overall would be great, until a user requires concurrent performance exceeding on chip RAM.

Hmm that's a good point! The thing I worry about is that the whole appeal of the Mac Pro is upgradeability — you can replace components over time. So integrated RAM would be problematic since that's a component people definitely like to upgrade. But with your idea... I dunno, if they could pull that off that would be super cool!

Yeah I think so too. If they can execute from off chip Ram, perhaps with a wait state or whatever it takes, for a ton of use cases no one will even notice.

It will all just effectively be large RAM.

Doing that coupled with a fast SSD, and people could be doing seriously large data work on relatively modest machines in terms of size and such.

A very simple division could be compute bound code ends up being on chip RAM, I/O bound code of any kind ends up in big RAM, off chip.

Doing just that would rock hard.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#514

Post by dang »

leptons wrote: post #470
leptons wrote: post #242

Earlier quoted context omitted.

Qualcomm chips don't have on-chip DRAM at the highest clock rate possible. If they did the same things Apple is doing in that way, the performance advantage of M1 shrinks. There's no real "magic" in the M1 or Apple, it's a reality distortion field that is hitting everyone right now. It's entirely possible to outperform the M1, there just hasn't been a need to do that until now - no major OS would have benefitted from…

Wow, downvoted by apple fanboi shitheads, without a single reply with any good counter arguments. "Hackernews" is a real dumpster fire, even worse than reddit.

If you continue to degrade the quality of this site even further, we will have to ban your account. We've asked you about this more than once already.

https://news.ycombinator.com/newsguidelines.html

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#515

Post by spronkey »

Bayart wrote: post #450
whizzter wrote: post #161

Earlier quoted context omitted.

Not to forget that apparently some of the people driving that scrappy team that had a big part in Intel's resurgence has been working on the M1. This actually reminded me of how I got skeptical comments from people when I told them that my small and scrappy Pentium-M (P-M -> M1, HAH!) based laptop was almost as fast as their desktop P4 monsters in compiling code.

I remember people using socket adapters to put Pentium M chips on their desktop. I also have very bad memories of my P4 Prescott around the same era. What an absolute trash pile of an architecture.

It wasn't actually that bad... Everyone likes to remember it as this massively power hungry beast with terrible performance. It was power hungry compared to what came later, but it did also provide some pretty good performance at the time - especially if you were the kind of tweaky nutter who liked to overclock.

Northwood was a relative bargain when you cranked the bus clocks up well beyond what it said on the box.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#516

Post by spronkey »

bydo wrote: post #345
rualca wrote: post #294

Earlier quoted context omitted.

I'm also curious to learn where was the bottleneck in both tests. I've read stuff about how the M1-packing MacBook Air shipped with a SSD whose burst write speed was far higher than the one shipped in older MacBooks. The bottleneck on build jobs tends to lie on disk access, specially with projects comprised of a significant number of small files whose build also outputs a bunch of small files. This is one of the reas…

The SSDs are faster, but only by ~20%: https://techcrunch.com/wp-content/uploads/2020/11/BlackMagic... From the TechCrunch review, which is pretty breathless but also contains a lot of good data: https://techcrunch.com/2020/11/17/yeah-apples-m1-macbook-pro...

Sequential performance is only one facet of overall SSD performance. Some older Apple SSDs weren't the greatest at random reads and writes - it's possible these new ones are quite a bit faster in that department, FWIW.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#517

Post by spronkey »

w0utert wrote: post #496
imtringued wrote: post #486

Earlier quoted context omitted.

It's in the same ballpark of power efficiency as AMD's x86 chips. It's slightly more efficient because of 7nm vs 5nm but if you scale it up to desktop frequencies it's going to consume the same amount of power as Ryzen CPUs.

Yes but the point is that Apple doesn't need to scale the M1 up to desktop frequencies, because it already is faster than x86 in single-threaded workloads, at lower clocks and significantly lower power. To scale up the multithreaded performance they just have to increase the core count and scale the cooling system accordingly, ie: exactly like you would have to for x86. A decent desktop cooler can dissipate enough he…

Multicore scaling on M1 doesn't appear to be as efficient as that on Ryzen - for example, the 4200U is able to beat the M1 in multicore tasks at quite a similar power draw, but gets soundly beaten in single core. Ryzen's big advantage over previous multicore implementations was infinity fabric - so clearly there's more to scaling than just the actual compute cores.

Don't forget that scaling up is also not just about frequencies, there are also packaging considerations - the CPU dies have to actually be able to dissipate the heat generated, and the package itself has to be able to do so as well. I'd expect that this is something AMD and especially Intel have a leg up over Apple with - although, considering they've already tread that ground it makes Apple's job a bit easier too.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#518

Post by spronkey »

vgchh wrote: post #88
cercatrova wrote: post #26

Is anyone else thinking, what the f*ck? Are we in a new era of computing? It certainly feels that way when looking at these desktop class ARM chips, where performance doubled every year or so, just like back in the 80s and 90s.

This is the Apple equivalent of a Tesla showing up and out competing much more expensive cars on the performance. It feels like a huge disruption and probably Apple’s chance to gain significant PC market share.

The issue with this is that Apple has the same problem that they have had for years - they don't really offer mainstream products.

Even their most inexpensive products are on the higher end of the pricing spectrum, and they could be 5x the performance but it still wouldn't matter.

There's a reason why Chromebooks are so popular and it sure isn't anything to do with their performance.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#519

Post by spronkey »

unsigner wrote: post #409
bob1029 wrote: post #378

Earlier quoted context omitted.

I thought this for a few seconds and then realized these sorts of benchmarks are lacking obvious controls (i.e. compiling for different architectures as noted by others). If you want a more realistic sense of how the M1 performs relative to x86 peers in raw, equivalent workloads, there are some cinebench numbers appearing out there: https://hexus.net/tech/news/cpu/146878-apple-m1-cinebench-r2... I am way more interes…

People are celebrating a kickass electric bicycle doubling the range of all the existing ones, and you demand to be shown it scaled up to the same power+price budget as an open pit mining truck, hauling dirt in three shifts. It might or might not happen, but the electric bicycle is still impressive.

Well, it's more like people celebrating an electric scooter that has twice the range of the electric bike.

Not saying it isn't a novel form of transport or equally useful in most cases, but... we're comparing a very stripped-down SoC to systems that have vastly more complexity for several different reasons - not least the ability to support more modular CPU/memory/GPU configurations.

Rework the existing x86 cores that we have into similar configurations and we'll likely see pretty substantial efficiency gains there too.

Re: 16-inch MBP 2x slower than M1 MacBook Air in a real-world Rust compile

#520

Post by spronkey »

The_Colonel wrote: post #298
harrygeez wrote: post #264

Earlier quoted context omitted.

My concern is not with performance but the seemingly energy consumption advantage. If Arm chips perform just as well (or as we've just seen, better) than x86 chips while sipping much less power, why would anyone want x86 on mobile devices? Are there any workloads that requires or perform better with x86?

This is clear only when you compare against an obsolete Intel which is known to be throttling power hog. The difference might be way smaller or non-existent once we see some comparisons with latest AMD offerings (especially when we have also CPUs with comparable process).

Yeah, Anandtech's dive revealed some interesting benchmark results - such as the Zen 2 4200U part being quite close in both performance and power consumption (a little higher) to the M1 in heavily multithreaded tasks, but getting stomped in both aspects in single-threaded ones.

The M1 isn't a tiny power-sipping mobile part - it sucks power down just like AMD's 15W TDP CPUs do. The efficiency gains Apple are getting here are likely to be the result of several factors, only one of which are the CPU cores themselves.

Post reply on HN