Apple M2 vs. AMD Rembrandt vs. Intel Alder Lake Linux Benchmarks

[bug77]

I understand people like to keep tabs on Apple's silicon. But I'm afraid the cold truth is it's pretty much the end of the line for Linux on Apple hardware. Too many custom things to support. As we are witnessing, even support for M1 isn't up to beta-software levels. And while there is the possibility expertise in the community will grow, I am not confident it will grow enough to keep up with Apple.

[andyprough]

But I'm afraid the cold truth is it's pretty much the end of the line for Linux on Apple hardware.

That's not a bad thing. The millions of work hours wasted on supporting inferior crapple hardware by GNU/Linux devs over the years has taken a lot away from what could have been accomplished otherwise.

Not that it's a zero sum game. I acknowledge that people work on what they want to work on. But I think it would also be fair to say that the return on investment has been extremely poor, and probably always will be.

[V1tol]

Apple Silicon is overrated

By Apple? Yes. But it is still the fastest consumer ARM CPU on the market.

the point here, why Asahi linux performe better in some tests?

Less overhead from OS, compiler differences?

[mdedetrich]

Michael,

It is well-understood that the Macbook Airs (both generations) are slower than most of the other Apple Silicon macs in heavy multithreaded performance due to them being fanless. The M2 macbook air seems even more noticeable in this regard due to the higher power consumption of the SoC than the M1.

Also, the storage in the base model M2 devices (so far) has been significantly slower than the M1 generation due to using half of the NAND chips for the base models (previously using 2, now only using 1). From what is known at this point, it seems that the problem is only with the base-storage of 256GB.

Anyway, overall impressive showing, even considering these.

Exactly, you are comparing a CPU that has no active cooling compared to other laptops which need fans unless you want them to thermal throttle ridiculously.

Exactly. Air M1 was too amazing (Apple wanted to ensure people will like the switch to ARM64), so Apple had to cripple Air M2 to make consistent steps in the next models line. M2 consumes a bit more power (like newer Intel/AMD/NVidia chips), but Apple also decided for weaker/cheaper passive cooling (and there's no active cooling naturally in the Air models).

The storage problem is well known and affects only the cheapest model. Such capacity is not usable for people, who need to do a real work, so they take the 512GB+ models anyway.

From what I heard its the GPU part of the M2 chip that is significantly faster than the M1, the CPU portion only has fairly minor performance improvements.

In any case I am very happy with my M1 14" pro, there isn't any real competition for that form factor against comparable laptops at least if you care about battery life/thermals.

[Linuxxx]

Oh that was unexpected, if the 6850 ran with: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: balanced - ACPI Profile: balanced.
then it seams to be even more efficient than M2?

To bad that the sensors aren't supported currently.

I'm pretty sure that if the benchmarks were run with the performance governor, than the efficiency gains would naturally be even greater.

[onlyLinuxLuvUBack]

Exactly, you are comparing a CPU that has no active cooling compared to other laptops which need fans unless you want them to thermal throttle ridiculously.



From what I heard its the GPU part of the M2 chip that is significantly faster than the M1, the CPU portion only has fairly minor performance improvements.

In any case I am very happy with my M1 14" pro, there isn't any real competition for that form factor against comparable laptops at least if you care about battery life/thermals.

But there isn't any software for apple arm except minesweeper ? why you need so much space ?

[stormcrow]

Apple Silicon is overrated

Not really. The base model MBA has known limitations both in the M1 and esp. M2 configurations. It's not a compute powerhouse. It's designed for mobile productivity. It doesn't have active cooling so it throttles often. In the M2 configuration at the base model the I/O is half as fast as the M1 model because it only has half the bandwidth (two 128 GB storage chips v. 1 256 GB storage chip on the M2).

You don't buy MBAs to do anything compute intensive, that's foolish. You buy it because its battery life and thermal profile will squash any x86 competitor without trying. I'm unsurprised by the benchmarks because of the above and because nearly everything is highly optimized for x86 hardware. I own an M1 MBP and I literally don't care about any of those benchmarks. It's a solid performer, does what I ask it to, and will keep doing long after any x86's battery bites the dust... all while sitting on my lap without burning it.

[stormcrow]

But there isn't any software for apple arm except minesweeper ? why you need so much space ?

Bull.

[bug77]

That's not a bad thing. The millions of work hours wasted on supporting inferior crapple hardware by GNU/Linux devs over the years has taken a lot away from what could have been accomplished otherwise.

Not that it's a zero sum game. I acknowledge that people work on what they want to work on. But I think it would also be fair to say that the return on investment has been extremely poor, and probably always will be.

I'm sure with the know-how they gained, the developers went on to build other useful things. So, far from a zero-sum game in my book.

[andyprough]

I'm sure with the know-how they gained, the developers went on to build other useful things. So, [far] from a zero-sum game in my book.

I agree, and that's why I said it's not a zero sum game. However, what do they have to show for it? Some laptops and desktops that kind of so-so work for the most part. Nothing ever approaching the performance of non-Apple hardware, since Apple routinely uses 2nd rate hardware by design.

As you basically said in your post above - for what? So that at the end of the line Apple could completely shut the door on those efforts? Something that any halfway knowledgeable observer could have predicted they would do years ago.

And one of the worst aspects is that we've known for years that Apple sits back and gobbles up open source contributions while giving little or nothing in return. So with the Linux dev efforts to run on Apple hardware - who has really benefited? You could make a pretty good case that Apple watched those efforts and time and time again said "thank you very much, I'll borrow that", and probably stole ideas from the Linux coders with no attribution.