The M2 Macbook Air reviews are here

[leadeater]

7 hours ago, rikitikitavi said:

It matters more when you often don’t use a backpack to carry it everywhere.

But these amounts of difference don't make that much difference, you missed the point. Like I said it's on a spectrum, within a range you're hardly noticing a difference which puts it firmly in personal choice rather than any empirical measure such as "I can only hold it for X amount of time vs Y amount of time". On any given day how long this is changes with the exact same device.

 

Both weight and size matters 1 million times more in my ladder example than walking around with it, go and climb up a ladder and free hand it with one hand then type with your other and see how long you can do it. Now tell me simply walking around with it for 15 minutes at a time matters more for this...

 

We've all been students in our lives, we've all walked around holding books or laptops, or both, so nobody needs this explained. Everyone has and will do it. Further this carries over to professional life in the workforce like going between meetings, which is by far less time and less frequent than student life.

 

"May arms are not made of wet toilet paper", to quote myself.

 

7 hours ago, rikitikitavi said:

dongles

Yea these can all go be thrown in an industrial shredder. Dedicated ports, as can and has been proven time and time again does not add significant weight or thickness to devices and it's pure aesthetics followed by simplification of PCB design. A universally adaptable port is of course very versatile and a welcome technology but not as a replacement to an ethernet port for example.

 

Dongles are just bad, almost always.

[leadeater]

6 hours ago, saltycaramel said:

The whole computer barely bigger than a ram DIMM, yet so powerful. 

It's like 4-6 SO-DIMMs in size....

 

Also damn, they took that PCB and put it through a compactor. That must have been wildly complicated PCB design.

[rikitikitavi]

1 hour ago, leadeater said:

But these amounts of difference don't make that much difference, you missed the point. Like I said it's on a spectrum, within a range you're hardly noticing a difference which puts it firmly in personal choice rather than any empirical measure such as "I can only hold it for X amount of time vs Y amount of time". On any given day how long this is changes with the exact same device.

Yes, a personal preference. What I tried to say was that there is a market for portable laptops that are thin-er and light-er and it doesn't matter if it won't make an actual difference on one's endurance.

 

PersonA:

- doesn't necessarily need the same array of ports as personB, if any

- doesn't want to pay for the extras that he doesn't need

- wants slim-er and light-er even if it is imperceivable by him or if it is fake-ish (touchbarMBP's sloping edges).

 

Sure it is a spectrum... like a guy who has a slim card/id wallet and a cellphone on him vs a guy who has his pockets bulging out with shitload of stuff for no practical reason.

 

1 hour ago, leadeater said:

Both weight and size matters 1 million times more in my ladder example than walking around with it, go and climb up a ladder and free hand it with one hand then type with your other and see how long you can do it. Now tell me simply walking around with it for 15 minutes at a time matters more for this...

Sure, but you need the functionality that device provides so you just endure that inconvenience, while the other person walking around in 15min bursts doesn't need the same functionality, so why should he get the same device? I heard very similar complaint several times, so at least 3-4 people think this is a legit thing lol

Extra little bulk hnere, extra little bulk there and in the end we get cumulative extra-extra - I forgot the old saying.

 

1 hour ago, leadeater said:

Dongles are just bad, almost always.

My point above + the occasions when that aforementioned user needs that specific port, most probably rarely.

 

* Just to make sure - we are not talking about docks that are at the desk and not carried around - because I'm strongly believe in usefulness and convenience of docks no matter if you have 1 or 20 port on your laptop. 

[hishnash]

23 hours ago, wseaton said:

desktops with power saving cores

Turns out you do in fact want a few small e cores in a desktop.  For a given transistor budget for even heavy workstation workflows it is a good idea to have some very small e-cores. These take up a very small amount of die area but since they have there own L1 and L2 cache when tasks run on them the perofmance cores L1 and L2 cache is un-effected.

The reason this is a big benefit is if your running a heavy task and the os/background app needs to do something if you just have big fat cores one (or more) of those cores is going to be interrupted and as the core switches to do the task for the background/os your working data that was in L1 or L2 will be evicted. Then that background task will finish (maybe within a few nano seconds) and you will go back to work but all of your working data is gone from L1 and being even L2 so now you need to wait... and wait and wait... for it to be loaded back in from ram and you need to then wait for the cpu to decode your instructions fill the out of order buffer etc to get up to full speed. a 1 nano second of background execution on your cpu core might well lead to unto 100 nano seconds of you not really getting any work done, and these interruptions happen really often (every few mili seconds). 

But if the system as a few e-cores it can use these for these background tasks. 4 e-cores + their small L2 cache takes up a lot les space than 1 perfomance core with its cache.  
 

This turns out that with a given Transistor budget you are always better of having a few e-cores in the mix if you're doing any intensive (more than 1ms second long) tasks since you likly can fit a few e-cores in and around the chip without really increasing the size by much as these cores can also run find with less IO bandwidth and higher IO latencies so you do not need to massively increase your frabirc to accomodate them. For context they typicly take up less die area than a single lane of PCI. 

[leadeater]

7 hours ago, rikitikitavi said:

Yes, a personal preference. What I tried to say was that there is a market for portable laptops that are thin-er and light-er and it doesn't matter if it won't make an actual difference on one's endurance.

I didn't say there wasn't a market, I said the arguments for why most people support thinner and lighter laptops almost always amount to personal choice and not that the weight difference or thickness actually has much tangible difference. I just don't like bad reasoning, especially when the wide support of it drives a bad trend that affects all future designs for the worse.

 

As to your similar example, death by a thousand cuts is also a thing. Losing small things here and there over time still amounts to something, until one day you're looking at most laptops having a couple of ports and being told to "just accept dongles".

[Obioban]

My plan is to wait for the 3nm chips and buy a 14" MacBook Pro, to replace my M1 Air... but, god damn it, this new air is nice. I want one, a lot.

[saltycaramel]

 

 

 

7300 in Cinebench r23 multi when in its stable long term throttled state, drawing ~10W.

 

The same score as a 2018 14nm mid-range desktop 6-core Intel i5-9600K at stock frequencies with proper desktop cooling, drawing ~70W during the same test.

 

Not bad to have that kind of power in a fanless slim laptop just 4 years later. If this is throttling, sign me up. 

[Obioban]

^in my estimation, they should have done the testing before disassembling the computer/cooling system, instead of after.

[hishnash]

On 7/17/2022 at 2:00 PM, Obioban said:

they should have done the testing before disassembling the computer/cooling system, instead of after.

So fare I have not seen a single channel do good testing.  If your testing thermal throttling then its important to do a few things:
1) Measure the ambient room air temp make sure you have a large enough room for the device to not really effect that temp.
2) do a lot of runs resetting the device to the same starting conditions for each run (at least 30)
3) do not report avg, use median if you want to provide a single number but better still would be to draw a box plot of violin plot
4) when saying if is 100 points faster in benchmark X instead say. "A T test confirms it is faster in benchmark X with a confidence of 95%" the median score is 100 points higher.  

All of the current stuff I have seen to date is people running a single run and them comparing the score to something else (not always even numbers captured on the same environment) and saying it is a few % faster without having any idea as to the margin of error. 

I did in the past have hopes that LTT labs would be able to provide this info (and im sure they will) but after what Linus said about audio and insisting on playing samples in videos (really useless) I don't think he wants to put in the effort to educate (its would be a lot of work) the users.  (he still shows AVG frame rate after all a completely useless number, should be showing distribution of frame devilry times and frame latency, in the high frame rate space frame latency is the most most important metric everything else is just a poor mans proxy for this why not show the real number). 

[Obioban]

Honestly, I think all the focus on benchmarking this laptop are a bit silly in the first place.  This is Apple's thin and light/ultraportable/consumer laptop. The fact that it can, say, edit four 4K streams at once at all is kind of absurd/unnecessary-- complaining that it does it 10 seconds slower than the computer with a fan seems relatively meaningless. 

[saltycaramel]

People tend to forget the kind of craptastic CPUs and limited use scope the 13" Airs and 12" Macbooks had from 2008 up to that glorious day of November 2020 when the best fanless laptop of all times (the M1 MBA) was released. 

 

And yet they were super popular for their intended use. 

 

Now we have Airs, both the M1 and the M2, that rival a mid-range desktop PC from 4 years ago. 
And yet we have tech influencers and tech bros "laptopsplaining" the Airs to regular normie users. 

- "you should care about the SSD drama"

- "you should care about the throttling" (which is called an "issue" or "problem" instead of intended behavior within Apple's "one chip, many thermal envelopes, superb burst performance everywhere" optimized silicon strategy)

- "you should care about spending 200$ more for a product you'll keep 5-7 years" (MKBHD is telling people to default to picking the 999$ M1 Air over the base 1200$ M2 Air because of the SSD drama and the price, imagine looking like a$$ in Zoom calls for the next 5 years on that last decade 720p webcam because you bought into the SSD-drama)

- "you should just give up the whole idea of the Air and go for the +30%_weight +30%_thickness 14" MBP"

 

Lots of Airsplaining and group-think ("everybody is making a big deal of the SSD drama so I should too") from tech influencers going on.

[Obioban]

All that said, I'm having a VERY hard time choosing between a new Air and 14" Pro-- not because of anything processor related, but because I can't decide if I care more about the size of the air or the screen/speakers of the Pro...

 

I have a bit of time to choose, as I'm going to wait out the 3nm chips either way (though there's been some rumors the M2 Pro/Max will be 3nm, so that might just push me sooner than expected). 

[rikitikitavi]

11 minutes ago, Obioban said:

All that said, I'm having a VERY hard time choosing between a new Air and 14" Pro-- not because of anything processor related, but because I can't decide if I care more about the size of the air or the screen/speakers of the Pro...

 

I have a bit of time to choose, as I'm going to wait out the 3nm chips either way (though there's been some rumors the M2 Pro/Max will be 3nm, so that might just push me sooner than expected). 

15” Air is rumoured.

 

Why would M2 family suddenly shrink?

[saltycaramel]

9 minutes ago, Obioban said:

All that said, I'm having a VERY hard time choosing between a new Air and 14" Pro-- not because of anything processor related, but because I can't decide if I care more about the size of the air or the screen/speakers of the Pro...

 

I have a bit of time to choose, as I'm going to wait out the 3nm chips either way (though there's been some rumors the M2 Pro/Max will be 3nm, so that might just push me sooner than expected). 

 

Gurman said Apple is internally targeting a fall 2022 release for the updated MBPs but they could be delayed to spring 2023. It's also high time to release the new Minis (M2 and M2 Pro). Would Apple debut the M2 Pro on the MacMini months before launching the M2 Pro/Max MBPs? Stranger things have happened. 

Yes the M2 Pro/Max/Ultra will be on the 3nm node. 

 

As for your dilemma, I feel you, I'll say one thing about the 14"/16" mini-LED screens: HDR and Dolby Vision content looks amazing on them, if you plan to also consume content regularly on the laptop. 

[saltycaramel]

3 minutes ago, rikitikitavi said:

Why would M2 family suddenly shrink?

 

Analyst Jeff Pu said he's hearing the M2 Pro/Max will be on TSMC 3nm node. 

 

Pro/Max/Ultra are their own family, all based around the Max. (cut to become a Pro, doubled to make an Ultra) 

[Obioban]

23 minutes ago, rikitikitavi said:

15” Air is rumoured.

 

Why would M2 family suddenly shrink?

Yeah, but a 15" Air will still have the Air level display/speakers. I'm not looking for bigger, I'm looking for better 

(Mini-LED/dolby vision) 

[Paul Thexton]

4 hours ago, saltycaramel said:

I'll say one thing about the 14"/16" mini-LED screens: HDR and Dolby Vision content looks amazing on them

After watching content on my 14” Pro and seeing what HDR is supposed to look like, I suddenly got offended when watching movies on my living room TV and found a deal on a 55” LG C1 ..

[DANK_AS_gay]

I don't know why he didn't put a thermal pad in. If you care about performance, and you bought a MacBook Air, you would stick a 2mm thermal pad in there. That way you get the better chassis, $400 cheaper for the desirable config, and the same performance. But ok, make up poor thermal tests that in no way represent real world workloads, and expect a fanless computer to survive it without throttling.

[Video Beagle]

20 hours ago, saltycaramel said:

- "you should care about the SSD drama"

What I really want to know is what the "slower SSD" actually means...like to real people.

the slower 256 gets a lot of air time because no click baiter has ever lost out by saying "apple bad", but in real world, real people using computers, not benchmark numbers, not weird percentages...was does it actually mean?

Give me comparisons of copying a 1 gig file and a 1 gig size folder full of images...that kind of thing that will actually start to replicated what a user sees. ( @GabenJrI'm calling to you on this)

[saltycaramel]

32 minutes ago, Video Beagle said:

What I really want to know is what the "slower SSD" actually means...like to real people.

the slower 256 gets a lot of air time because no click baiter has ever lost out by saying "apple bad", but in real world, real people using computers, not benchmark numbers, not weird percentages...was does it actually mean?

Give me comparisons of copying a 1 gig file and a 1 gig size folder full of images...that kind of thing that will actually start to replicated what a user sees. ( @GabenJrI'm calling to you on this)

 

The theory is that under some circumstances you would experience the effects of having a slower swap memory. 

 

In practice most users are better off anyway with the base M2 Air compared to the M1 Air or, god forbid, to any Wintel laptop in this form factor. 

 

[hishnash]

1 hour ago, saltycaramel said:

The theory is that under some circumstances you would experience the effects of having a slower swap memory. 

 

yer if you have only 8GB of ram and the entry level SSD in particular since the SSD controller shares the RAM with the system (normally a key advantage for these systems in latency) when you are under high memory pressure that means the controler needs to flush writes to the SSD much more oftener (as it cant cache as many of them in RAM) and needs to pull many more read ops. Added to this once the system starts to swap all of these swap evens need to be fully flushed otherwise they do not free up any RAM, and when you read you are reading from raw NAND not the DRAM buffer of the drive.

So 1/2 the bandwidth to the drive is most noticeable when you're under high memory pressure both in reads and in writes. 

In my few if it was hard to get enough 128 GB dies (as I suspect it was as 2 128 GB NAND dies is typicly cheaper, in normal times, than 1 256 GB die) they should have just started with 512.

 

[saltycaramel]

I'd rather see these new Airs gradually creep back towards that magical 999$ price point in the coming years than for them to default at half a terabyte of flash.

[Obioban]

49 minutes ago, saltycaramel said:

I'd rather see these new Airs gradually creep back towards that magical 999$ price point in the coming years than for them to default at half a terabyte of flash.

Inflation being what it is, I'm not confident we'll ever get back there.

[rikitikitavi]

8 hours ago, saltycaramel said:

I'd rather see these new Airs gradually creep back towards that magical 999$ price point in the coming years than for them to default at half a terabyte of flash.

Would be nice to have M1 Air become full-fledged SE-tier device with a much lower entry price.

[GhostRoadieBL]

It's a little cliche at this point but lack of repair-ability is a tough con to get past, the pro's are stacking up really well compared to basically everything else on the market even at the price point.

considering the XPS 13 is not the best windows machine for the biggest macbook air competitor, dell just isn't innovating anymore (neither is Apple when looking at the rest of the laptop compared to M1 or earlier)

I guess we'll be waiting to see the trickle down 1000nit miniLED displays from Apple and some kind of QD-OLED from dell to really get into DCI-P3 consumer displays on smaller workstation laptops. Then it'll be the battery wars for media consumption and on the go work. It's probably going to be an interesting few years if companies start to compete again for recession dollars instead of just tossing a new chip in the same old chassis and screen combos. 16in macbook air with 99Wh would be epic!