I don’t really like geekbench. It always feels like it favors apple devices.
Sometimes companies build product to get high scores, rather than to work well in real settings.
Apple has done a remarkable job with their silicon though.
I got an used M1 2020 MacBook Air (base model, no upgrades) back in 2023 and that thing is still going strong. Doesn’t feel slow at all. Not even slightly.
I bought an M1 MacBook Air as my first Apple Silicon Mac when they hit the market (longtime Mac user, so direct comparison to Intel Macs). It blew me away. Still have it, despite replacing it with a MBP. At this point, it’s the extra screen off to the side playing videos while I nerd out in my office, though it can do much more.
And can do that for two seconds before throttling.
Don’t worry guys, we’ll get a RISC-V device that blows Apple ARM out of the water any second now.
Annnnyyy second now…
They always do this comparison and it is always misleading. They are the first to get a processor on the new TSMC node, so of course the processor will be faster.
Okay, but that Intel CPU was part of the first family of CPUs on Intel’s equivalent node (18A) too. If nothing else, it serves as an interesting Intel vs TSMC lithography comparison. Which one would expect TSMC to win anyway, but at least now we have numbers.
They’re the first to get the fastest chips, so saying their chips are the fastest is… misleading? But also “of course” it’s true?
I’m confused.
I don’t think anybody was misled into thinking it meant this chip would always be the fastest on the market.
Saying the first new CPU on a new node is fastest is uninteresting from a technical standpoint, as there isn’t any competition yet, even if it is not inaccurate.
The average consumer doesn’t care why it’s the fastest. For a time, they’ll have a faster chip than their peers. Only people on Lemmy would say “yeah but that’s only because new nm process, so it doesn’t count!”
If the conversation is between chip manufacturers talking shop, that matters. For everyone else, it’s irrelevant.
but can it run linux and real software? what would I want more powerful cpu for if not linux server?
Apple silicon still only supports Apple’s proprietary operating systems but you can run Linux in a virtual machine or use Asahi which is still not recommended for daily use.
I like my iPhone but damn it’s locked down.
I bought it, I should be able to go to GitHub, grab a project, compile it and run it.
I’m surprised that so many technically-adept people find this acceptable.
Not necessarily the bit about compiling software directly on the phone; I can do that and I don’t, but the part where Apple has the last word on what apps can exist (yes, I know there are some very narrow escape hatches).
As a developer I routinely compile projects that Apple has never seen and run them on my iOS devices. It’s not locked down in the sense that it’s impossible, but that the barrier is too high for normal users (who also don’t have a developer program membership).
Im Europe we have alternative app stores on iOS too.
But at the end of the day, even as a technically adept person, I don’t care that much about what my phone can do beyond the actual phone things, and just want it to be a reasonable smartphone that doesn’t need me to be technically adept just to be useable.
Are beowolf clusters still a thing? If yes, imagine one made out of those.
Yes, using the Mac Mini.
All of this performance is wasted on Apple’s closed hardware and OS. What a shame.
It’s also unfortunate timing. The most valued chips these days are the ones with high parallelism, for AI performance. If Apple turned its eye to GPUs, maybe they could challenge Nvidia.
Their iGPUs have serious performance, but it doesn’t matter because they insist on stupid proprietary APIs that no real developer bothers to write software for.
If we’re talking about AI performance the software is there, and models often get ported to MLX.
On the iPhone, sure, but it would be equally wasted on Android phones because who needs this much performance in a phone anyway. However, I expect them to also use it in next year’s Macbook Neo, which is a surprisingly capable laptop for its price with even the A18 Pro chip. It also bodes well for the next generation of their M series chips if the A20 Pro is this fast… and there are TONS of people using Macs for real work. Code compilation in particular got super fast with the M1 Pro and I believe it’s great for creative applications too, though I haven’t really used those much myself. Creatives types gravitated towards MacOS even before Apple Silicon, especially video editors because of Final Cut Pro.
I believe the power can be used with Android’s Desktop Mode, or even Android’s current multitasking capabilities like multi-window and pop-up window.
More cats in my dwarf fortress are never wasted.
Honestly, DF was the last time i thought about single core performance.
I think Minecraft is single-threaded
Hasn’t really been for years, and minecraft isn’t a terribly hard game to run anymore anyway.
With the new vibrant visuals and ray tracing features on bedrock, midrange devices cannot keep up.
Those are both GPU-dependent, and certainly not single-threaded.
I’m over here on a 15 year old computer.
it could be an apples-to-oranges comparison. not sure if that intel core can be inside apple phones.
IIRC an A20 is an SOC rather than a CPU too.
Intel cpus are SOCs too.
Soc is almost a meaningless term these days. It used to mean “a mobile CPU”. But for power reasons laptop CPUs have all become SOCs too.
Technically these are all System in Package (SiP) rather than systems on a chip (SoCs). It’s not cost effective to waste silicon area on things like memory and storage in the latest and greatest TSMC nodes, so they have several different silicon dies that they connect together using advanced packaging techniques, but where each silicon die in that package might come from a different foundry/process.
The comparison that matters is always left out. Computing performance by dollar cost.
Even higher, since those Apple silicon chips are very efficient.
You misread. OP mentioned price/performance, and I think you’re referring to performance/watt.
Oh, you mean the upfront price per chip, not the operating cost. Funny enough, the A20 chips are no more expensive (and possibly cheaper) than their Intel and AMD counterparts. The main issue is that the A20 is locked inside an iPhone, which is overpriced.
It does matter but performance per currency alone is also kinda shit of a metric too. Might point you towards getting like 50 atom CPUs but then good luck finding memory for all of them to do anything useful.
Better to show performance between each and then list the cost to let the user decide if it is worth it. Make your charts between hardware in set price ranges is also pretty good.
not really apples to apples (no pun intended) comparison. Apple has designed their silicon chips with the intention of running their operating systems and the OS is certainly optimized for the architecture. Whereas x86 is somewhat generic and the OS performance varies, Windows and Linux not necessarily returning the same results. Linux can be highly optimized, but not Windows.
No pair of different CPU models is an apples to apples comparison. But comparing two CPUs of the same ISA with similar implementations using tact frequencies and core numbers and cache sizes might make sense. Like two AMD Ryzen CPUs or two Intel CPUs.
OK, I mean, there are benchmarks for this kind of thing.
It’s honestly amazing what these chips can do.
Unfortunately it’s stuck behind the Apple ecosystem.
Linux is inevitable
Slop
Apple will fight it with everything they can. Latest MacOS27 hides linux partitions from the bootpicker so it is blocked in firmware.
Not exactly, the bootpicker just won’t show partitions that are not flagged as bootable. This is a departure from previous versions, that would check to see whether partitions were bootable without looking at the flag, and took a little bit of investigating to figure out this undocumented behavior, but it is an easy fix:
Evil fucks
The phone has been out for 10 days and this article is from 6 days before it was released.
Why are we looking at an “early benchmark” from before release instead of an actual real benchmark of the device. I don’t get why Geekbench doesn’t have results for iPhone 18, or why this “early benchmark” calls it “iPhone19,2.”
If true, the single core performance of this chip is almost identical to the single-core performance of the new M6 chip.
Something doesn’t seem right, but I don’t know a lot about Geekbench or iOS benchmarks to know if this seems legit or not.
If true, the single core performance of this chip is almost identical to the single-core performance of the new M6 chip.
The two “super” cores on the A20 pro are the exact same as the two on the M6 asside from some extra L2 Cache on the M6.
Source: https://youtu.be/t8bxkCKNiuA
Do Linux distros run on Apple hardware? The hardware seems faster than intel/amd but I hate MacOS.
Asahi kinda works, but only up to m2/m3. Apple isn’t very helpful with Linux support.
For Apple silicon, Asahi Linux is the most committed to that, they seem to be able to get things going on Mac but always a few generations behind.
Pretty good if you buy a Mac second hand tbh, Apple hardware is still the best in durability in my experience.
I don’t think there’s a distro for running Linux on any recent iPhone, which is what this chip is for.
Asahi Linux is the leading distribution for Macs, but it’s a few generations behind with hardware support. M3 support is in the works but still experimental and not fully featured yet.
M3 was moved to offical support this month!
