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

[mdedetrich]

That it's perfectly fine to compare passive vs active chips. Just keep in mind that one is silent and the other not.

No its not, there is nothing scientific about. If you want to do this correctly you need to minimize the variables and establish some control/s, thats how scientific method works.

What about comparing cars, is it ok to compare a Porsche 993 vs McLaren F1?

That example is both not apt and stupid

And where is your proof for that? Have you read this article and the ones preceding it? Because the 6850U is faster than the M2 at roughly the same power consumption.

The fact that there isn't a single fanless x86 laptop that exists?

Again why is a x86 CPU that consumes 17 W at average and 28 W peak (CPU only) (https://www.phoronix.com/review/ryzen7-6850u-acpi) not passiv coolable but the M2 with 25 W and 50 W peak (at wall) is?
And just because one is cooled with another cooling solution you're not allowed to compare them? Doesn't make sense to me.

Because you don't understand how the scientific method works. Even the reported power consumption isn't can't be directly correlated because Apple's CPU is a SOC (system on chip) which means the CPU/GPU/memory/SSD are all soldered into a single chip so even power reporting isn't apples to apples (literally and pun intended). Likewise doing it from the wall is pointless because know you are not just comparing CPU's but everything in the laptop (i.e. Apple laptops tend to have much nicer displays that hence draw more power because they can go up to 1.2k nits iirc)

Is it ok to compare 2 active Laptops where one has a stronger fan? And if so why?

From least stupid to most stupid ideally you would have the same chassis with the same fans with the same fan profiles. Since that is extremely difficult the next best thing is to normalize for fan profiles. The absolute stupidest thing is to compare laptops with no fans with laptops that have fans and even stupider than that is comparing a thin and light (which is what an Macbook Air is) to laptops that are more catered to power users (i.e. thicker 14"+ laptops).

At this point you may as well compare phones to a god damn laptop and it would make just about as much sense.

[Anux]

Secondly there are multiple reasons on the ISA/silicon level why x86 is worse, x86 is an old ISA

As is ARM.

that is meant to maintain compatibility with programs that are programmed/built 40 years ago. All of that extra compatibility has to be accounted for in the silicon where as the way ARM works is it redesigns the ISA every generation (which means no old cruft).

No ARMv8.5 (the base of Apples M chips) is backwards compatible to ARMv5. Apple left AArch32 away wich is not allowed in the ARM standard (AArch32 becomes optional in ARMv9) so you could actually say M1/2 is not an ARM compatible chip.
Also x86 left some compatibility away over the years, see recently 3Dnow on Zen.

all instructions have the same fixed word length

Thats basically every RISC arch.

The very test whose results we're discussing demonstrates three state-of-the art products and the M2 apparently wins the performance-per-watt game.

Are you talking about this phoronix article under which we are writing right now? Can you please help me see how you came to that conclusion?

No its not, there is nothing scientific about. If you want to do this correctly you need to minimize the variables and establish some control/s, thats how scientific method works.

So one can also not compare supercomputers with different cooling solutions despite all scientists that work with them do it?

That example is both not apt and stupid

Why? They have different cooling solutions and slightly different driving properties.

The fact that there isn't a single fanless x86 laptop that exists?

From were do you get your "facts"? Passive Laptops were a new thing maybe 10 years ago there are 100s of models to this day. Let me give you a short introduction to researching facts. You type 'google.com' in your web browser, then there is a search field in which you type 'fanless laptop' and hit enter.
Thats even worse than your "thats not science, you're not allowed to compare that.

Because you don't understand how the scientific method works.

I'm not sure if I have to take that serious from a guy that graps "facts" out of his ass.

Even the reported power consumption isn't can't be directly correlated because Apple's CPU is a SOC (system on chip) which means the CPU/GPU/memory/SSD are all soldered into a single chip so even power reporting isn't apples to apples

You can easily compare the powerconsumption at the wall. You have a point that we are missing that for the AMD system in this case. But the fact that we have a difference of 5 W is enough for the rest of the system to be powered. Maybe if we search hard enough, the internet might also spit out real messured values (haven't done that yet).

Apple laptops tend to have much nicer displays that hence draw more power because they can go up to 1.2k nits iirc)

You can't regulate the brightness on Apple hardware? Can't believe that.

From least stupid to most stupid ideally you would have the same chassis with the same fans with the same fan profiles. Since that is extremely difficult the next best thing is to normalize for fan profiles. The absolute stupidest thing is to compare laptops with no fans with laptops that have fans and even stupider than that is comparing a thin and light (which is what an Macbook Air is) to laptops that are more catered to power users (i.e. thicker 14"+ laptops).

Nothing stupid there if you don't care for those properties.

At this point you may as well compare phones to a god damn laptop and it would make just about as much sense.

Yes nothing wrong there, modern smart phones are easily usable as simple desktops (also you can connect them with screens and keyboards) so why not compare them with a laptop? Just make it clear at the beginning what and why you compare it.
Actually you can compare every computing device that is capable of supporting your workload with another. One might not be usable mobile like a PC the next would have a small screen like a phone.

[Dukenukemx]

That's not the argument people are making. People point out that a fanless laptop with much lower power budget has a very competitive performance against machines that require active cooling to provide the same performance.

I keep linking Louis Rossmann who lost his mind how Apple didn't put a fan on the CPU but on the other end of the laptop, essentially making the fan useless. Not an M1 or M2 but an Intel laptop. Apple hasn't been cooling their Intel chips for a while now, and it ended up burning out a number of them.

You may dislike Apple however you want but the M1 is a commendable leap forward in terms of power efficiency.

No it isn't, and none of you pro Apple supporters look at benchmarks. You all take this rhetoric and run away with it without knowing anything. The M1 uses the same amount of power in a multithreaded work load as competitive x86 laptops, but is only more efficient in single threaded work loads. Given 4x more efficient in single threaded but still. If you include the GPU for gaming then it's horrible. People have posted videos here showing this.

[drakonas777]

Look how many "CPU experts" dreams were shattered with actual facts. Imagine getting this destroyed still against products on inferior process nodes.

x86 >>>>>> ARM.

Good one LOL

A lot of people are telling a lot of truth there. Problem is they tackle many "orthogonal" facts/aspects/properties and some ideas get lost in that. Personally, i think Apple Silicon is amazing as a piece of hardware. I'd like something similar to emerge from Qualcomm/Nuvia, NVIDIA or else on the PC. The more hardware diversity the more interesting computers world. I'm just a bit allergic to the hype and misrepresentations regarding M chips. That's all.

Oh, and as for "hot and loud" x86 notebooks - for starters factory system/SoC/CPU power settings could be not a idiotic garbage they are now usually, which contributes to the myth that x86 has garbage efficiency. Same goes for desktop parts. In the desktop at least you can control it, while some notebooks does not allow to fine grain the power. For example, I personally would always limit power to the least TDP down value. Performance drops are negligible (on ryzen at least, ha ha), but power savings are phenomenal. I switched my 3700X to 45W now and It's cool and quite, much more than RPI4 sitting next to it in deskpi pro case with active cooler LOL

Anyways, it was a nice discussion. In 2 years I will do that testing myself - plenty macbooks in the work office in artists department (fronted programmers i mean LOL)

[Raka555]

Apple hasn't been cooling their Intel chips for a while now, and it ended up burning out a number of them.

That is one of the reasons I never wanted to have an Intel based MacBook.

No it isn't, and none of you pro Apple supporters look at benchmarks. You all take this rhetoric and run away with it without knowing anything. The M1 uses the same amount of power in a multithreaded work load as competitive x86 laptops, but is only more efficient in single threaded work loads. Given 4x more efficient in single threaded but still. If you include the GPU for gaming then it's horrible. People have posted videos here showing this.

I don't care about the benchmarks.
What is important to me is that I get more than one day of real world use out of a single charge of my MacBook M1-Pro, while developing and compiling software, at speeds that exceed what a Ryzen 9 5900x can do.

No throttling, no noise, no heat, no 1/2 speed when on battery. Nothing else can match it at this point.

What you see in your benchmarks are probably some synthetic torture test that does not resemble the real world usage.

[smitty3268]

This already exists with TSMC 5nm. We have M1 which uses TSMC 5nnm and AMD'z Zen3+ which is TSMC 5nm which laptops are starting to use. Why do we have to wait for 5nm (i.e. the renamed N5), are you arguing there is going to be some magic with N5 that will benefit x86/64 more compared to ARM in some future theoretical product?

That's incorrect.

The latest Zen3+ AMD cpus are TSMC 6nm, which in spite of seeming very close to 5nm is simply a refinement of the significantly older 7nm process. Zen 4 will be the first AMD chip running on TSMC 5nm, and you'd need to wait for laptop chips based off that (probably a year away?) to get a fair comparison in terms of manufacturing.

That said, I think the whole concept of trying to "equalize" laptop chips is silly. Every case/system design varies wildly in terms of heat management and the simple fact is that Apple spends more money in order to get the newest manufacturing capabilities before it's competition. You have to keep that in mind if you are trying to compare architectures, but ultimately who cares? It's the end products that should be compared, not some theoretical thing nobody can actually compare because there are a million different factors complicating things.

So that leaves us with the actual products, like this article tested. And yeah, it would have certainly been nice if Michael had included something like a M1 Pro as well, but at the end of the day I imagine he only has so much money to spend on hardware since Apple isn't likely to be shipping him any to review. It makes sense for him to try and get a M2 chip to test, since he already had an M1 product around.

[Anux]

A lot of people are telling a lot of truth there. Problem is they tackle many "orthogonal" facts/aspects/properties and some ideas get lost in that. Personally, i think Apple Silicon is amazing as a piece of hardware. I'd like something similar to emerge from Qualcomm/Nuvia, NVIDIA or else on the PC. The more hardware diversity the more interesting computers world. I'm just a bit allergic to the hype and misrepresentations regarding M chips. That's all.

I do totally agree with that.

Fact is, right now you can't by a passive laptop that is as powerful and efficient as the current M1/2. If silent and performant is your most important demand, you have no competition in the near sight.
Also I think with a 6850U you could build a similar device, maybe not as thin and slower at video editing (and other hardware accelerated stuff in M1/2) but roughly the same performance.
The 6850U is worse at idle and probably the whole platform is because you need an (LP)DDR interface and PCIe support but that mainly reduces the battery runtime with light task to maybe 12h vs 18?

[drakonas777]

I do totally agree with that.

Fact is, right now you can't by a passive laptop that is as powerful and efficient as the current M1/2. If silent and performant is your most important demand, you have no competition in the near sight.
Also I think with a 6850U you could build a similar device, maybe not as thin and slower at video editing (and other hardware accelerated stuff in M1/2) but roughly the same performance.
The 6850U is worse at idle and probably the whole platform is because you need an (LP)DDR interface and PCIe support but that mainly reduces the battery runtime with light task to maybe 12h vs 18?

And I do totally agree with that. I understand the advantages of ARM ISA as well as the advantages of M1/2 SoC/system design in contract to current Intel/AMD "APUs" and systems. However, I do believe that Intel/AMD will throw some interesting stuff in the next few years and with some tinkering around power settings that stuff can potentially be very good too. We'll see

[Anux]

Yes it will be interesting indeed, both intel and AMD will integrate FPGAs in their CPUs and I wonder if this will profit the average customer in the future (make hardware video codecs able to be updated by software or become an AI accelerator at demand) or if this will be a server thing only.

[Raka555]

Yes it will be interesting indeed, both intel and AMD will integrate FPGAs in their CPUs and I wonder if this will profit the average customer in the future (make hardware video codecs able to be updated by software or become an AI accelerator at demand) or if this will be a server thing only.

Or when it becomes self aware, to improve its own processor