MBA M2 SSD read/write speeds slower than MBA M1

[Spline]

Just got my new MBA M2 and was surprised to find that the SSD read/write scores are lower than the M1, and that's with the 512 GB SSD, not the base 256 GB:

	​	Read​	Write​
OLD	MBA M1 (8‑Core CPU / 8‑Core GPU), 8 GB, 512 GB SSD	3,536 MB/s​	2,920 MB/s​
NEW	MBA M2 (8‑Core CPU / 10‑Core GPU), 24 GB, 512 GB SSD	3,271 MB/s​	2,647 MB/s​

Tested using Amorphous Disk Mark

On Blackmagic Disk Speed test, the M2 scored: Read: 2,935 Write: 2,336 MB/s

(I didn't run Blackmagic on the M1 before shipping it to the person I sold it to.)

FileVault was active on both Macs.

 

[Apple_Robert]

Translation = not going to notice with normal use.

 

[Slartibart]

🥱 … sorry… just woke up… something of importance happened? Otherwise, trust @Apple_Robert… at least on this one. 🤓

 

[Spline]

Translation = not going to notice with normal use.

That goes without saying.

As the performance has gone backwards, not forwards, however, it's always worth noting what the Apple hype machine obscures.

 

[mi7chy]

Slower but not as bad as 256GB.

 

[fakestrawberryflavor]

Translation = not going to notice with normal use.

While this is true, it's also true that we live in a world and time where future iterations of technology should not regress in any form or function from previous iterations, especially those coming in at a higher price point.

 

[kschendel]

While this is true, it's also true that we live in a world and time where future iterations of technology should not regress in any form or function from previous iterations, especially those coming in at a higher price point.

An interesting viewpoint, but I don't know how defensible it is when looking at the big picture. You're assuming that costs keep dropping, and that happens to not be true. For all we know, without the cheaper SSD arrangement (which has no effect on usable performance), the new M2 MBA might be even more expensive than it is. (Costs are more than component costs.)

 

[Apple_Robert]

While this is true, it's also true that we live in a world and time where future iterations of technology should not regress in any form or function from previous iterations, especially those coming in at a higher price point.

In a lot of ways, I think many people on MR have expectations that exceed the ability of current tech or exceed the expectation of the device, as to intent and purpose by the manufacturer. Too many here are doing nothing more than going in circles debating meaningless tests and benches. And many of those same people as well as onlookers can't make a simple decision anymore for all of the needless hyperbole, bias, and loss of focus. I don't know how some of these people make it out of bed in the morning given that the mental paralysis they have inflicted upon themselves.

A lot of times, I would agree with your general assessment. With the M2 MBA, I am not concerned and I don't think the average consumer who isn't obsessed with benches is concerned either. Given the M2 MBA doesn't have a fan, it very well could be that Apple has tweaked some of the hardware to prevent too much heat or, it could simply be a matter of Apple getting these parts at a good price and keeping their margins high. I am not going to presume to know the answer.

Overall, I have been thrilled with my M1 MBA (outside of the speaker grills). It was my favorite laptop of all time until I got the M2. For me, the new form and enhancements were worth upgrading. I think that is the bottom line for each person who buys the M2. Is the buyer happy after 1 or 14 days? Is the M2 meeting that person's needs day in and day out regardless of what YouTube says, tech site reviews or forums like MR? If the answer is yes, that is all the truly matters.

 

[jav6454]

Just got my new MBA M2 and was surprised to find that the SSD read/write scores are lower than the M1, and that's with the 512 GB SSD, not the base 256 GB:

	​	Read​	Write​
OLD	MBA M1 (8‑Core CPU / 8‑Core GPU), 8 GB, 512 GB SSD	3,536 MB/s​	2,920 MB/s​
NEW	MBA M2 (8‑Core CPU / 10‑Core GPU), 24 GB, 512 GB SSD	3,271 MB/s​	2,647 MB/s​

Tested using Amorphous Disk Mark

On Blackmagic Disk Speed test, the M2 scored: Read: 2,935 Write: 2,336 MB/s

(I didn't run Blackmagic on the M1 before shipping it to the person I sold it to.)

FileVault was active on both Macs.

Those are sequential speeds which you will never see in the real world. Test the Random 4K @ QD1 speed then get back to us.

 

[Spline]

Those are sequential speeds which you will never see in the real world. Test the Random 4K @ QD1 speed then get back to us.

Still lower on the M2 by a bit:

Amorphous RND4K QD1: M2 Read 59.05, write 36.58, M1: Read 63.09, write 39.42

 

[Slartibart]

Still lower on the M2 by a bit:

Amorphous RND4K QD1: M2 Read 59.05, write 36.58, M1: Read 63.09, write 39.42

yeah, that milliseconds in perception will make a real difference for us humans… especially if you consider the time a visual stimulus takes to be registered from hitting your retina to be processed and acknowledged by your brain. But still, nice measurement. 🤓

 

[Sydde]

… consider the time a visual stimulus takes to be registered from hitting your retina to be processed and acknowledged …

Right, but if you are copying/transferring really large / a lot of files, the difference can become noticeable. Fortunately, Apple will be leveraging the Neural Engine to address this issue, programming it to study your behavior and begin what you want done before you even realize you want it done, thus obfuscating performance deficiencies behind it already being done.

 

[jav6454]

Still lower on the M2 by a bit:

Amorphous RND4K QD1: M2 Read 59.05, write 36.58, M1: Read 63.09, write 39.42

That is within 3MB/s read/write difference. If anything, it will add 2-3 extra seconds loading times.

 

[Slartibart]

Right, but if you are copying/transferring really large / a lot of files, the difference can become noticeable.

How? First and foremost you have no reference… neither visual nor mental. Secondly - as others and some research points out - this difference is within the variance of measurements. yes, there is a difference- but you have to apply geological timescales to end up with something that is of perceptive relevance 🤜🤛🤓