M2 MacBook Air throttling is being vastly over-exaggerated

[unrigestered]

It is basically the performance you had in iPads for a long time already. It is nothing new.

how long is "for a long time"?
the A12Z in the 2020 iPad Pro has a performance under sustained load in Cinebench more in the the half the speed of the M1 range. In Geekbench the differences are slightly less pronounced.
The M1 iPad Pros came after the MBA where the speeds should indeed be quite close for most things.

That's not to say that the A Series iPad CPUs were / are not great, but the whole experience is different, since the OS is way more restrictive and most apps are cut down in features too, compared to what you could get on "fully fledged computers"
Not that everything of these need to be necessary for everyone, but i'd call the OS and apps themselves more powerful as well.

 

[Warped9]

Should keep the M1, skip the M2, M3 and wait for M4 - M5 Machines tbh.
M2 isnt really a upgrade.

Skip the M5. Too much chance it might be unstable and start destroying things and killing people.


*Now how many will get this?

 

[adrianlondon]

Skip the M5. Too much chance it might be unstable and start destroying things and killing people.

But that's what I'm buying it for.

 

[russell_314]

Skip the M5. Too much chance it might be unstable and start destroying things and killing people.


*Now how many will get this?

I hear the M7 should be able to wirelessly connect to the Death Star. If you're into destroying planets that might be a useful feature

 

[tornadowrangler]

The M1 Air also throttled.
Here is the evidence:
View attachment 2033864

Compared to M2 Air's throttle:
View attachment 2033865
Last year not a lot of people notice or talk about the M1 Air's throttling due to the crazy gains from intel. But from here on out there will not be a jump like that again until Apple jumps to the next architecture, maybe RISC-V which is at least 10+ years away I think.

Ok cool. I didn't say it M1 MacBook Air didn't throttle. I said it wasn't seen as an issue. The perception is that the M2 Macbook Air throttles more in some way than the M1 Macbook Air.

I was just making the point that if the M1 had throttled more, than we WOULDN'T see it as a problem that the M2 does. Whether or not it is true, someone found a way to get people to think that the M2 throttles more (which I DO NOT see as a problem).

The point I was trying to make was to NOT compare the M1 and M2, but compare the M2 in the new design Macbook Air to what it would do in the old design. If it would throttle less, even a little less, in the old design, then people would see it as a step backward. BUT if the M1 had come out in the new design and throttled just as much, people WOULD NOT have seen it as a step back.

It is a good example of the anchoring bias.

 

[Dnzilla]

I agree, but I’m betting that the refreshed mbp, whenever it comes, will see a price increase.
The timing is just a case of apple feeling like they can get away with it.

Inflation hitting everything and apple won’t escape it..

 

[DavidChoux]

Thank you! That app is a good one! I will look into it.

Does your 2016 MBP not turn off turbo boost in low power mode in MacOS Monterey?

I tried the low power mode but still saw my CPU over 2GHz (base speed) so for some reason it doesn't work.

 

[Toratek]

Skip the M5. Too much chance it might be unstable and start destroying things and killing people.


*Now how many will get this?

So the question is, whose engrams were used for the circuits? 😆

 

[Zest28]

how long is "for a long time"?
the A12Z in the 2020 iPad Pro has a performance under sustained load in Cinebench more in the the half the speed of the M1 range. In Geekbench the differences are slightly less pronounced.
The M1 iPad Pros came after the MBA where the speeds should indeed be quite close for most things.

That's not to say that the A Series iPad CPUs were / are not great, but the whole experience is different, since the OS is way more restrictive and most apps are cut down in features too, compared to what you could get on "fully fledged computers"
Not that everything of these need to be necessary for everyone, but i'd call the OS and apps themselves more powerful as well.

The M1 chip is basically the A14X. So we have had the "M1" chip for ages in the iPad.

The experience isn't different, people didn't notice a performance difference between the 2018 iPad Pro and the 2021 iPad, as it is simply the A12Z/X vs the A14X / M1.

It is only in "heavy tasks" such as gaming where the difference between the M1/A14X iPad Pro and the A12Z / A12X iPad started to show.

Basically what Apple did was put the iPad chip in the Mac.

 

[nicho]

The M1 chip is basically the A14X. So we have had the "M1" chip for ages in the iPad.

Why is M1 in quotes? We've actually had the M1 chip in iPads for ages...

 

[unrigestered]

so someone was using a duo, or quad core PC for a long time, basically had a 64 core, or an Intel based supercomputer with 10.000s of CPUs, since they share the same technology?

 

[Zest28]

so someone was using a duo, or quad core PC for a long time, basically had a 64 core, or an Intel based supercomputer with 10.000s of CPUs, since they share the same technology?

The A12Z is a 8-core CPU with 4 efficiency cores and 4 high performance cores using the cores from the iPhone. Does that sound familiar to you?

It is why the M1 can be put in the iPad Air even, as it is basically the unreleased A14X.

 

[PsykX]

The real issue is the slow performance with the 256GB model.
Take the 512 and it's a perfect computer.

 

[Schizoid]

Got an M2 Macbook Air day 1. 16gb and 1TB. Thought I’d put it through its paces not with benchmark software, but with apps I’d actually run…
Office, Safari, Chrome, Photoshop, Keynote , Mail etc - didn’t even flinch running all.
Final Cut Pro, rendered a 4K 6 min vid, super quick compared to my old Intel mac. Yes it got hot and throttled slightly, but one thing I noticed is that once it finished rendering, it dissipated the heat VERY quickly, we’re talking seconds. I think Apple engineering knows what they’re doing!
My ultimate test was running a few games connected to an external monitor, it got hot pretty quickly, but I left it running for a good hour or so. Still peformed admirably and again dissipated the heat extremely quickly at the end and of the session.
One thing I did notice was the windowServer process seemed to grab 100% of the cpu after connecting an external monitor. Never seemed to relinquish the CPU until I restarted. This is a separate issue though, more macOS I believe.

 

[1BadManVan]

Got an M2 Macbook Air day 1. 16gb and 1TB. Thought I’d put it through its paces not with benchmark software, but with apps I’d actually run…
Office, Safari, Chrome, Photoshop, Keynote , Mail etc - didn’t even flinch running all.
Final Cut Pro, rendered a 4K 6 min vid, super quick compared to my old Intel mac. Yes it got hot and throttled slightly, but one thing I noticed is that once it finished rendering, it dissipated the heat VERY quickly, we’re talking seconds. I think Apple engineering knows what they’re doing!
My ultimate test was running a few games connected to an external monitor, it got hot pretty quickly, but I left it running for a good hour or so. Still peformed admirably and again dissipated the heat extremely quickly at the end and of the session.
One thing I did notice was the windowServer process seemed to grab 100% of the cpu after connecting an external monitor. Never seemed to relinquish the CPU until I restarted. This is a separate issue though, more macOS I believe.

Can’t wait for my 16/1tb to show up. Overkill for my needs but wanted to treat myself with my first ever mac lol

 

[Gelam]

I tried the low power mode but still saw my CPU over 2GHz (base speed) so for some reason it doesn't work.

Interesting, I thought the Haswell and Broadwell generations were the only ones without the function.

 

[theotherphil]

Yep, looks like the throttling is really hurting the M2 MBA. Totally smashed an Alderlake i7 XPS13 costing $250 more.

/s

 

[Zest28]

Yep, looks like the throttling is really hurting the M2 MBA. Totally smashed an Alderlake i7 XPS13 costing $250 more.

/s

Is that Max Tech doing benchmarks again? Where is Rene and all the Apple defenders to stop these toxic benchmarks?

Oh wait, the M2 MBA looks good in these benchmarks, then I guess it is alright.

 

[thenewperson]

Is that Max Tech doing benchmarks again? Where is Rene and all the Apple defenders to stop these toxic benchmarks?

Oh wait, the M2 MBA looks good in these benchmarks, then I guess it is alright.

Yeah, it's good.

 

[Zest28]

Yeah, it's good.

Btw, he should actually include AMD systems. Dave2D did a similar test against the M2 MBA, and the closest to the M2 MBA was the Zenbook.

 

[LeeW]

I didn't say it didn't throttle

Yes, you did.

I think the main problem is that the M1 MacBook Air didn't have this throttling problem

 

[thenewperson]

Btw, he should actually include AMD systems. Dave2D did a similar test against the M2 MBA, and the closest to the M2 MBA was the Zenbook.

Cinebench 🤢

 

[jdb8167]

Is that Max Tech doing benchmarks again? Where is Rene and all the Apple defenders to stop these toxic benchmarks?

Oh wait, the M2 MBA looks good in these benchmarks, then I guess it is alright.

Not really. I only had time to watch the first half but it was pretty clear that he didn't really decide what he wanted to test. First it starts out as a battery test starting at 60% and in the XPS in the lowest power mode. Then he switches to the highest power mode. Then he notes that the performance of the Alder Lake SoC is limited when on battery. I don't know if eventually he tested with a power adapter but this was not a scientifically valid test. Choose your settings and stick with them. Then change things and try again. Then note the differences.

It's not better just because the MacBook Air is better on battery. Either do a single test or do something comprehensive but just randomly changing things isn't very helpful.

 

[1BadManVan]

Not really. I only had time to watch the first half but it was pretty clear that he didn't really decide what he wanted to test. First it starts out as a battery test starting at 60% and in the XPS in the lowest power mode. Then he switches to the highest power mode. Then he notes that the performance of the Alder Lake SoC is limited when on battery. I don't know if eventually he tested with a power adapter but this was not a scientifically valid test. Choose your settings and stick with them. Then change things and try again. Then note the differences.

It's not better just because the MacBook Air is better on battery. Either do a single test or do something comprehensive but just randomly changing things isn't very helpful.

Maybe watch the entire video? The Xps died in the middle of the Geekbench compute test and he had to plug it in. The m2 still had 40% left. And he only briefly switched power modes to show the performance difference between the modes. The remainder of the test he had to run mostly plugged in since the battery was toast

 

[diamond.g]

Not really. I only had time to watch the first half but it was pretty clear that he didn't really decide what he wanted to test. First it starts out as a battery test starting at 60% and in the XPS in the lowest power mode. Then he switches to the highest power mode. Then he notes that the performance of the Alder Lake SoC is limited when on battery. I don't know if eventually he tested with a power adapter but this was not a scientifically valid test. Choose your settings and stick with them. Then change things and try again. Then note the differences.

It's not better just because the MacBook Air is better on battery. Either do a single test or do something comprehensive but just randomly changing things isn't very helpful.

They ran out of battery power on the XPS box so the remainder of the tests were done with the system plugged in.

Edit: Beated by @1BadManVan