Simple SSD Question

[Jenkon]

Hi,

Can someone please confirm this for me…

If I have Samsung T7 external SSD’s that transfer at 1000mb/s, regardless of if I get a base MBA M1 (with a read/write I’d say 2000mb/s) or a base MBA M2 (with a read/write is 1500mb/s) I’m still only going to get a transfer speed of the T7 at 1000mb/s, correct?

So the whole MBA M2 SSD read/write issue is redundant in that specific use case?

Thx.

 

[Slartibart]

Theoretically: Yes. Basically it’s physics. Not magic. 😎

Practically: you won’t get the specified max. transfer speed for the T7 on M1 or M2. Expect a reduction by 10-25%, occasionally 50%.

External SSDs in a Thunderbolt 3/4 enclosure that supports up to 40gbps of theoretical max speed will be deliver much higher transfer speeds.

The T7 is limited by the USB 3.1 Gen 2 support Apple implemented.

This isn’t limited to AS- but Intel-based Macs too: see here, here or here.

IMHO it’s still a really nice user experience to use a Samsung T7 (or T5 for that matter). YMMV. 🤓

 

[Jenkon]

So in theory, the M1 or M2 would make no performance difference here. It’s just down to the external SSD speed?

 

[Slartibart]

So in theory, the M1 or M2 would make no performance difference here.

In relation to TB and USB the technical specification published by Apple seem to be indeed the same for the M1 MBP and the M2 MBA.

It’s just down to the external SSD speed?

Strictly theoretical: yes.

If you have a receiver which supports a higher speed then the connected transmitter offers, transfer will saturate at the max speed the transmitter is able to deliver. Theoretically.

 

[Fishrrman]

You won't get 1000MBps speeds from any t7 regardless of which Mac you connect it to.

You'll more likely see reads in the 900-930 range, and writes a little lower.

 

[w5jck]

I don't have a Samsung T7, but I do have a Samsung T5. Here are the speeds I get with 2 Macs and an iPad:

MacBook Air M1 2020 = 357 MB/s (calculated by copying a 8 GB video file)
MacBook Air M1 2020 = 357 MB/s (calculated using BlackMagic)

MacBook Pro 2014 = 357 MB/s (calculated by copying a 8 GB video file)
MacBook Pro 2014 = 448 MB/s (calculated using ATTO Disk Benchmark)

iPad Air 2020 = 195 MB/s (calculated by copying a 8 GB video file)

Note: The T5 is rated as USB-C Gen 1 up to 5 Gbps (625 MB/s). So I get a little over 50% of the rated speed on both Macs. Elderly Intel Mac was same speed as MacBook Air M1.

 

[jwahaus]

I have 3's Samsung T7's (the original, the T7 Touch, and the T7 Shield). I've benchmarked them on both an M1 Mini and a M1Pro MBP

M1 Mini - 725MB/s Write and 662MB/s Read
MBP - 790MB/s Write and 720MB/s Read

Even though in theory the 10Gb/s USB 3.1 interface should be enough to get the full 1000MB/s from the SSD drive there is a considerable amount of overhead that limits the speed. The only way to get faster external SSD speeds is to use a costly Thunderbolt connected drive.

 

[Jenkon]

I have 3's Samsung T7's (the original, the T7 Touch, and the T7 Shield). I've benchmarked them on both an M1 Mini and a M1Pro MBP

M1 Mini - 725MB/s Write and 662MB/s Read
MBP - 790MB/s Write and 720MB/s Read

Even though in theory the 10Gb/s USB 3.1 interface should be enough to get the full 1000MB/s from the SSD drive there is a considerable amount of overhead that limits the speed. The only way to get faster external SSD speeds is to use a costly Thunderbolt connected drive.

So the whole issue of the SSD being slower on the MBA M2 vs the MBA M1 is redundant in the use case scenario?

 

[jz0309]

So the whole issue of the SSD being slower on the MBA M2 vs the MBA M1 is redundant in the use case scenario?

I don’t know what you mean by “redundant”.
The INTERNAL 256 GB SSD ind the M2 MBA is slower than the INTERNAL 256GB SSD in the M1 MBA.
External speed will the similar to each other

 

[Jenkon]

I don’t know what you mean by “redundant”.
The INTERNAL 256 GB SSD ind the M2 MBA is slower than the INTERNAL 256GB SSD in the M1 MBA.
External speed will the similar to each other

As in the speed is dictated by the slowest component in the chain, in this case the T7 external dive, as far as the transfer rates are concerned. So it wouldn’t matter if you were using the M1 or M2.

 

[Isamilis]

As in the speed is dictated by the slowest component in the chain, in this case the T7 external dive, as far as the transfer rates are concerned. So it wouldn’t matter if you were using the M1 or M2.

If you copy to/from internal SSD then yes, transfer rate will be limited to the internal SSD speed (the lowest). This is my opinion only, as I don’t have M2 to test.

 

[Jenkon]

But in the case the T7 would be the slowest as the M2 has a read/write of approx 1500mb/s compared to the T7 1000mb/s

 

[BanditoB]

The SSD speed slowdown that's been noted on the M2 MacBook Airs is due to the use of one flash memory chip on the motherboard instead of using two chips. This means that the SSD controller is unable to read/write to two chips in parallel which doubles the throughput or speed. With only one chip to read/write to, the performance is half of what you would get with two chips. This only applies to the internal SSD because of this one-chip configuration. The SSD controller that runs the internal SSD is still capable of much higher speeds if configured correctly.

That being said, the internal SSD controller has no impact on the read/write speeds of external drives. They have their own controllers and configurations. 99% of the time, it's the interface between the external SSD controller and the computer it's connected to that is the limiting factor for read/write speeds. USB is generally slower that Thunderbolt, but has caught up with the most recent specifications.

Long story short, the slower SSD speed of the MacBook's internal SSD has no effect on external drive speeds, which it sounds like was your concern.

 

[Jenkon]

The SSD speed slowdown that's been noted on the M2 MacBook Airs is due to the use of one flash memory chip on the motherboard instead of using two chips. This means that the SSD controller is unable to read/write to two chips in parallel which doubles the throughput or speed. With only one chip to read/write to, the performance is half of what you would get with two chips. This only applies to the internal SSD because of this one-chip configuration. The SSD controller that runs the internal SSD is still capable of much higher speeds if configured correctly.

That being said, the internal SSD controller has no impact on the read/write speeds of external drives. They have their own controllers and configurations. 99% of the time, it's the interface between the external SSD controller and the computer it's connected to that is the limiting factor for read/write speeds. USB is generally slower that Thunderbolt, but has caught up with the most recent specifications.

Long story short, the slower SSD speed of the MacBook's internal SSD has no effect on external drive speeds, which it sounds like was your concern.

Thank you for your thorough answer. Yes, this is what I was seeking clarity on. No point fussing about SSDgate when the only activity I would notice the slower read/write speed (transferring files) is dictated by my external hard drives speed, and not if I choose an M2 or M2 machine (which both have higher internal SSD’s than the external SSD’s).

 

[Fishrrman]

OP:

If you buy a t7, and get read speeds faster than around 950MBps, let us know.
(doesn't matter WHICH Mac you connect it to...)

 

[jav6454]

Hi,

Can someone please confirm this for me…

If I have Samsung T7 external SSD’s that transfer at 1000mb/s, regardless of if I get a base MBA M1 (with a read/write I’d say 2000mb/s) or a base MBA M2 (with a read/write is 1500mb/s) I’m still only going to get a transfer speed of the T7 at 1000mb/s, correct?

So the whole MBA M2 SSD read/write issue is redundant in that specific use case?

Thx.

The bottleneck is always the bus with the slowest interface. So yes, regardless of which Mac you get, the bottleneck will be that Samsung SSD when doing Sequential read/writes.

Random? You’ll never see a bottleneck.