Do M1 MacBooks suffer from the same performance issues the Mac mini M1 does with USB-C

[syrcular]

Hi all,



I've recently discovered through some testing that the Mac mini M1 seems to have slow performing USB-C/Thunderbolt 3/4 ports than my intel based MacBook Pro. When testing fast external SSD drives, I've noticed that my intel based 16 inch MacBook Pro will typically perform at about 900MB/s when testing the disk speed with read and write, while the Mac mini M1 using the same drive connected with the same cable will only perform at about 500-600MB/s in read and write speeds. This is a significant performance difference and I've seen others report this. My question as I prepare to possible purchase a secondary MacBook Air or MacBook Pro; Does any of the M1 notebooks suffer from similar performance issues with connected USB-C SSD drives or is this issue isolated to the Mac mini M1?

 

[HatMine]

From what I’ve seen, there is no difference. But you should be able to work around this by using, for example, a thunderbolt 3 dock with 10 Gbit ports, since the heart of the issue is that some USB-C devices connect at a speed of 5 Gbit directly to the Mac. Whether this is a bug or not, I can’t say.

 

[syrcular]

From what I’ve seen, there is no difference. But you should be able to work around this by using, for example, a thunderbolt 3 dock with 10 Gbit ports, since the heart of the issue is that some USB-C devices connect at a speed of 5 Gbit directly to the Mac. Whether this is a bug or not, I can’t say.

Unfortunately since a Thunderbolt dock uses the effected USB or Thunderbolt port, it effects the overall speed of the connected dock. I was able to confirm this issue using an OWC Thunderbolt Dock.

 

[HatMine]

Unfortunately since a Thunderbolt dock uses the effected USB or Thunderbolt port, it effects the overall speed of the connected dock. I was able to confirm this issue using an OWC Thunderbolt Dock.

What do you mean? The speed should be higher through the dock since it has its own USB controller, thus bypassing the “5 Gbit USB issue” of the port, since the port is talking to the dock with thunderbolt running at 40 Gbit. This video explores this (iirc)

 

[syrcular]

I have that exact dock and my tests are showing even slower read write speeds to the external SSD when connected to it vs connected directly into the Mac Mini.

 

[mj_]

In theory there shouldn't be any performance difference between an M1 Mac and an Intel Mac when it comes to external SSDs. The fact that you're experiencing a massive slowdown with the same SSD in the same case using the same cable when connected to an M1 Mac indicates that there's something else wrong. Could be some sort of hardware or software incompatibility. Let's address the latter first: is the filesystem on the drive in question encrypted with FileVault 2? What file system, HFS or APFS (or FAT32 or ExFAT)?

 

[leman]

I have seen reports that Apples USB controller is sometimes slower than Intel controllers. Unfortunately, this seems to be a hardware issue since a software bug would have probably been solved already. This is basically the first revision of Apples own high-performance USB interface, and it focuses on safety and security first and foremost, so I think the performance will improve in future hardware iterations. But current M1 owners are probably stuck with marginally slower USB.

 

[syrcular]

In theory there shouldn't be any performance difference between an M1 Mac and an Intel Mac when it comes to external SSDs. The fact that you're experiencing a massive slowdown with the same SSD in the same case using the same cable when connected to an M1 Mac indicates that there's something else wrong. Could be some sort of hardware or software incompatibility. Let's address the latter first: is the filesystem on the drive in question encrypted with FileVault 2? What file system, HFS or APFS (or FAT32 or ExFAT)?

So I'm currently testing between two different computers.

16 inch MacBook Pro 16 and a Mac mini M1

I'm also testing two different high speed external SSD's

Samsung T7 and G-Drive SSD (Latest Generation)

Both SSD's are pre-formatted with HFS+ (Journaled)

I'm using the included cable for both SSD's, assuming they are both 10gb cables (will be able to confirm this with my MacBook intel test - see below)

When testing the Samsung and G-Tech SSD's on my MacBook Pro Intel connected directly to one of it's Thunderbolt 3 ports, I get roughly about 900-950GB/s Transfer speed

When testing the Samsung and G-Tech SSD's on my Mac mini M1 connected directly to one of the Thunderbolt ports I get roughly 500-600GB/s Transfer speed

When connecting an OWC Thunderbolt 4 hub to my Mac mini M1, and connect the Samsung or G-Tech SSD's to the Thunderbolt 4 port, I get roughly 400-500GB/s

As you can see from the results above, it's clear that this is not file system related, but perhaps a hardware issue. The intel clearly performs at the expected speeds, while there is a noticeable drop in performance when testing all of this through the Mac mini M1.

I just watched that MaxTech video that the other poster here posted, but I'm not getting the same results. Presumably this could be an issue with my Mac mini, but I've read others are having the same issue.

 

[ADGrant]

Hi all,



I've recently discovered through some testing that the Mac mini M1 seems to have slow performing USB-C/Thunderbolt 3/4 ports than my intel based MacBook Pro. When testing fast external SSD drives, I've noticed that my intel based 16 inch MacBook Pro will typically perform at about 900MB/s when testing the disk speed with read and write, while the Mac mini M1 using the same drive connected with the same cable will only perform at about 500-600MB/s in read and write speeds. This is a significant performance difference and I've seen others report this. My question as I prepare to possible purchase a secondary MacBook Air or MacBook Pro; Does any of the M1 notebooks suffer from similar performance issues with connected USB-C SSD drives or is this issue isolated to the Mac mini M1?

This is a known issue with M1 Macs. Their USB-C/Thunderbolt performance is significantly weaker than the Intel Macs.

Interestingly, the M1 iPad's performance is significantly worse. There is a YouTube channel that covers these issues extensively. Their latest video with times is below.

 

[Mike Boreham]

Same issue as this thread?

https://forums.macrumors.com/thread...ally-10gb-s-also-definitely-not-usb4.2269777/

 

[bobcomer]

It's the same hardware, so a Macbook anything wont be any better. I think it's a weakness in the thunderbolt hardware itself. We see it with slower than normal ethernet adapters and I think the external video problem some of us see (interference of a kind) is also related, but that may be a GPU problem. Remember it is V1 hardware and shows it...

 

[commentzilla]

I have that exact dock and my tests are showing even slower read write speeds to the external SSD when connected to it vs connected directly into the Mac Mini.

Yep, it's true. You are not imagining it. I have been complaining about this issue since the 1st TB4 dock it the market. It affects all of the TB4 docks. I had both a OWC and CalDigit TB4 dock and they exhibit the same issue. You will also see this issue to a much lesser extent with the TB3 docks. I owned a OWC and Akito TB3 dock too.

This issue is documented on OWC's website in their blog. Wait until you read what the solution is. what they don't make obvious is that with a TB4 dock, you will often get speeds SLOWER than a TB3 dock. I returned all of my TB4 docks because my TB3 docks are faster!

In order to get the fastest write speed of an external drive when used in conjunction with an M1 Mac, the following conditions are required:

1. The storage device (either bus-powered or AC “self-powered”) MUST be connected directly to one of the M1 Mac’s Thunderbolt ports.
2. Plus ONE of the following:
   • Connect a TB3 Display to the second Thunderbolt port on the M1 Mac.
   • Connect a TB3 Dock such as our 14-Port Dock, which offers native multiple display type support (HDMI, DisplayPort, or VGA without needing an active adapter) to the second Thunderbolt port on the M1 Mac. In this scenario, no external display needs to be connected to the M1 Mac.
   • Connect a TB3 device to the second Thunderbolt port on the M1 Mac ANDconnect a TB3 OR USB-C display to a Thunderbolt port on the TB3 device. The good news here is our Thunderbolt 3 Pro Dock and the Mercury Elite Pro Dock (a unique combination of storage and connectivity) both provide a speed boost. When using an OWC Envoy Pro EX (Thunderbolt 3) drive connected to an M1 Mac, a USB-C display connected to the OWC Thunderbolt 3 Pro Dock, and the Dock connected to the M1 Mac, there was a 56% write speed increase!

MORE WAYS TO GET FASTER EXTERNAL DRIVE SPEED WITH YOUR M1 MAC​

More Ways to Get Faster External Drive Speed With Your M1 Mac

In order to get the fastest write speed of an external drive when used in conjunction with an M1 Mac, the following conditions are required:

eshop.macsales.com

 

[commentzilla]

From the comment. Read OWCs response, which admits that ALL TB4 docks are hobbled. But hey that's ok because now you have more slower TB/USB ports! Ha.

More Ways to Get Faster External Drive Speed With Your M1 Mac

In order to get the fastest write speed of an external drive when used in conjunction with an M1 Mac, the following conditions are required:

eshop.macsales.com

 

[commentzilla]

If you have an M1 Mac. You're better off with a TB3 dock unless Apple comes up with some kind of fix, which I doubt.

 

[Fomalhaut]

So I'm currently testing between two different computers.

16 inch MacBook Pro 16 and a Mac mini M1

I'm also testing two different high speed external SSD's

Samsung T7 and G-Drive SSD (Latest Generation)

Both SSD's are pre-formatted with HFS+ (Journaled)

I'm using the included cable for both SSD's, assuming they are both 10gb cables (will be able to confirm this with my MacBook intel test - see below)

When testing the Samsung and G-Tech SSD's on my MacBook Pro Intel connected directly to one of it's Thunderbolt 3 ports, I get roughly about 900-950GB/s Transfer speed

When testing the Samsung and G-Tech SSD's on my Mac mini M1 connected directly to one of the Thunderbolt ports I get roughly 500-600GB/s Transfer speed

When connecting an OWC Thunderbolt 4 hub to my Mac mini M1, and connect the Samsung or G-Tech SSD's to the Thunderbolt 4 port, I get roughly 400-500GB/s

As you can see from the results above, it's clear that this is not file system related, but perhaps a hardware issue. The intel clearly performs at the expected speeds, while there is a noticeable drop in performance when testing all of this through the Mac mini M1.

I just watched that MaxTech video that the other poster here posted, but I'm not getting the same results. Presumably this could be an issue with my Mac mini, but I've read others are having the same issue.

I have exactly the same computers, but use a CalDigit TB3 hub.

Like you, I get about 30% less disk speed when connected directly to the Mac Mini via a 10Gbps USB connection. I get about 80-150MBps more (Samsung T7) when connected to the 10Gbps USB port on the CalDigit hub - so I'm glad I bought the CalDigit and not the OWC hub! The T7 is still about 150-200MBps slower than on the Intel Mac, even through the TB3 hub.

There is a whole thread on this subject at: USB on M1 Macs isn't actually 10Gb/s? (Also definitely not USB4)

The M1 Mini also has lower USB connection speeds with some external enclosures. I have at least one SSD that connects at 10Gbps on my Intel MBP16 and only 5Gbps on the M1 Mini. It does connect at 10Gbps on the CalDigit hub, and transfer speeds are similar to the MBP16 - maybe because the SSD is SATA3 and the bottleneck is the SSD itself, rather than the interface.

Interestingly, even though a 5Gbps USB connection should in theory get close to 500MBbp (assuming 10/8bit encoding), the SATA 3 SSD only hovers around the 300-330MBps mark, when it is close to or exceeds 500MBps when connected via 10Gbps USB. There is a lot more overhead on these connectors than marketing would suggest. If you take total bitrate of interface, 5Gbps / 8 = 625MBps, but you actually get roughly 50% of that when not contrained by the bandwidth of the storage medium. In short, for USB, choose an interface twice the bandwidth you actually want from your storage to avoid being bottlenecked by the interface.

 

[marksthompson]

What do you mean? The speed should be higher through the dock since it has its own USB controller, thus bypassing the “5 Gbit USB issue” of the port, since the port is talking to the dock with thunderbolt running at 40 Gbit. This video explores this (iirc)

I have just bought the OWC Thunderbolt 3 hub (5 port). USB 3.0 devices work fine (HDD 100-150Mbs, ssd 300 Mbs). When i plugged in a Crucial X8 USB 3.1 Gen 2 drive throughput plummeted to 150 Mbs when plugged in hub from 750 Mbs when plugged directly into my Macmini M1. I do not have any other higher speed devices to check at the moment. Both my supplier and I contacted OWC and to be honest, their responses where dismissive and uncaring. They admitted there was a problem, blamed the M1 Mac controllers and did not appear to understand why they should remove the claims on their website that the hub works with M1 Macs. I am sending the hub back. Hugely disappointed with OWC's attitude. The problem might be with the Mac M1 but that does not absolve OWC from warning their customers.

 

[Mike Boreham]

When i plugged in a Crucial X8 USB 3.1 Gen 2 drive throughput plummeted to 150 Mbs when plugged in hub from 750 Mbs when plugged directly into my Macmini M1.

Are you sure you are just seeing the write cache effect? Those numbers are close to the change that happens after the cache is full, in this review:

https://www.anandtech.com/show/1618...6-and-x8-2tb-review-qlc-for-storage-onthego/6

 

[marksthompson]

Mike, I do not know. Unless the benchmarking app (Blackmagic) is flooding the cache. Have you any suggestions o how I might test that theory?

 

[Mike Boreham]

Mike, I do not know. Unless the benchmarking app (Blackmagic) is flooding the cache. Have you any suggestions o how I might test that theory?

I am not expert at benchmarking drives but know my own efforts have sometimes been confused by write cache. Off the cuff I suggest leaving the drive connected for a while to let the write cache clear and then test again. Don't know how long is needed.

I have seen higher benchmarks with T5 and T7 drives connected to an OWC TB4 hub than direct to M1 MBA but not by the amount you saw.

I have also found AmorphousDiskmark to give more consistent results than Blackmagic. Free in the App Store.

 

[marksthompson]

Mike,
thanks for comments and suggestions. Will try both out.

 

[marksthompson]

Just a comment on re-reading your post. I am getting significantly worse throughput through the OWC hub than when directly connected to the TB ports on the Macmini M1. We are talking 180 MBs through the hub against nearly 800 MBs when directly connected.

 

[Mike Boreham]

Mike,
thanks for comments and suggestions. Will try both out.

Here is my back to back test of a Samsung T5 connected via the OWC hub and direct to M1 MBA.

 

[Mike Boreham]

Just a comment on re-reading your post. I am getting significantly worse throughput through the OWC hub than when directly connected to the TB ports on the Macmini M1. We are talking 180 MBs through the hub against nearly 800 MBs when directly connected.

I know! that's why I commented previously. I have never seen that kind of reduction for being direct connected and your numbers look very much like the drop that happens when the write cache is full. If you repeat that big drop I am out of ideas!

 

[BeatCrazy]

Yep, it's true. You are not imagining it. I have been complaining about this issue since the 1st TB4 dock it the market. It affects all of the TB4 docks. I had both a OWC and CalDigit TB4 dock and they exhibit the same issue. You will also see this issue to a much lesser extent with the TB3 docks. I owned a OWC and Akito TB3 dock too.

This issue is documented on OWC's website in their blog. Wait until you read what the solution is. what they don't make obvious is that with a TB4 dock, you will often get speeds SLOWER than a TB3 dock. I returned all of my TB4 docks because my TB3 docks are faster!

In order to get the fastest write speed of an external drive when used in conjunction with an M1 Mac, the following conditions are required:

1. The storage device (either bus-powered or AC “self-powered”) MUST be connected directly to one of the M1 Mac’s Thunderbolt ports.
2. Plus ONE of the following:
   • Connect a TB3 Display to the second Thunderbolt port on the M1 Mac.
   • Connect a TB3 Dock such as our 14-Port Dock, which offers native multiple display type support (HDMI, DisplayPort, or VGA without needing an active adapter) to the second Thunderbolt port on the M1 Mac. In this scenario, no external display needs to be connected to the M1 Mac.
   • Connect a TB3 device to the second Thunderbolt port on the M1 Mac ANDconnect a TB3 OR USB-C display to a Thunderbolt port on the TB3 device. The good news here is our Thunderbolt 3 Pro Dock and the Mercury Elite Pro Dock (a unique combination of storage and connectivity) both provide a speed boost. When using an OWC Envoy Pro EX (Thunderbolt 3) drive connected to an M1 Mac, a USB-C display connected to the OWC Thunderbolt 3 Pro Dock, and the Dock connected to the M1 Mac, there was a 56% write speed increase!

MORE WAYS TO GET FASTER EXTERNAL DRIVE SPEED WITH YOUR M1 MAC​

More Ways to Get Faster External Drive Speed With Your M1 Mac

In order to get the fastest write speed of an external drive when used in conjunction with an M1 Mac, the following conditions are required:

eshop.macsales.com

Are these "conditions" above applicable to just USB devices? Because I can get full TB3 speeds (2700MB/s read/write) for SSDs connected both directly to my M1 mini, and connected via the CalDigit TB4 Element Hub.

I'm using the ACASIS TB3/USB enclosure, with a 1TB WD_BLACK SN750.

 

[Mike Boreham]

Here is my back to back test of a Samsung T5 connected via the OWC hub and direct to M1 MBA.

View attachment 1829516

In the above test the Samsung T5 (256GB) was faster connected via the OWC (four port TB4) hub, than connected direct to my M1 MBA.

In the test of my Crucial X6 (4TB) below it is the other way round. The X6 is faster when connected direct to the M1 than via the hub.

I have bouts of doing this testing and end up getting confused trying to see patterns and consistency. For me all my drives are plenty fast enough direct or via the OWC and differences are not noticeable without measurement.

I guess if I was seeing down to 180MB/s I would be concerned.