USB on M1 Macs isn't actually 10Gb/s? (Also definitely not USB4)

[albercook]

Disk drives are absolutely not rated in bits per second, where are you getting this from? Except if you mean USB drives, but the Gbps in that case attempts to clarify what type of USB3 port is supported rather than indicating the speed of a drive. For example, USB 3.1 Gen 2 = USB 3.2 Gen 2x1 = SuperSpeed USB 10Gbps = USB 3.1 10GBps, with the last being currently most common in marketing materials. You can buy drives that support 10Gbps USB3 that vary in speeds from <100MB/s to slightly over 1GB/s, which will saturate the port. The whole point of this thread is that the drive Gbps ratings are especially useless on M1 Macs, since many drives will only connect with slower USB protocols than is supported, or will operate at significantly slower speeds than on PCs. As you've even seen yourself, when you searched for the rating of the internal drive, it was rated in MB/s not Mbps, which made your whole post trying to convert the external drive speeds from bytes to bits for comparison purposes completely unnecessary. Literally every single drive speed measurement tool and all drive marketing materials or specifications I have ever seen (and I've seen many, from many different companies), measure speed in bytes, not bits.

And to answer your question:
The absolute maximum data bus bandwidth of a TB3/4 port is 22Gbps, which can be reduced down to 8Gbps if the same cable/TB controller is used to connect to monitors. Therefore the absolutely maximum possible data transfer speed to a device connected to a TB3/4 port is 22Gbps/8=2.75GB/s or 2750MB/s. As you found out yourself, this is rather close to speeds of MBA's internal SSD and more generally, the rated speeds of currently most popular good PCIe 3.0 x4 NVMe SSDs. So yes, with a high quality TB3 enclosure and a good NVMe drive inside it, you will be able to have your external drive reach read speeds close to 90-100% and write speeds 75-90% of those of thee internal drive.

I did start with "...I'm not sure that I'm reading this correctly..." No surprise that I was wrong. So should I have said bus bandwidth is in Gbps and the values in BlackMagic are in MB/s. When trying to understand the performance issues discussed in this thread I found that confusing.

Thank you for the clear answer to my question.

I'm still not sure how to read what adonis3k said: "For the price/performance the Orico TB3 40GB/s is the best IMO....with 2TB Sabrent Rocket.."

If I understand adonisk3 should have said "Orico TB3 40Gbps instead of "Orico TB3 40GB/s"

That is why I was converting adonis3k's BlackMagic speeds were 1819/2229 MB/s. to 18Gbps

Finally, is adonis3k saying that the combination of Orico TB3 and Sabrent Rocket is one of the combinations that achieve 70 - 100% of the internal ssd specifically on the M1?

Thank you for helping me understand this thread better.

 

[adonis3k]

What I mean is for the price compared to the other external enclosures I have tried this was the fastest one I got, your numbers may vary depending on the drive/size you may use.

This is a speed test on my mac mini m1 from the internal drive 256GB...which is very similar to the Orico external drive I have plugged in...

 

[Mike Boreham]

I have not read the whole thread, but chipping in to contribute my experience with an M1 MBA, a 24" LG Ultrafine Display (which has 2 x TB3 ports and 3 x USB-C 5Gbps ports), and three Samsung T5 drives (2TB, 500GB and 250GB).

My results are summarised on this table (all MB/s):

Main Conclusions:

The M1 TB3 port is ~30% slower than the LG Display TB3 port.
The M1 TB3 port is ~9% slower than the 5Gbps USB ports on the LG Display.

I repeated the Direct-to-M1 tests with and without the LG Display plugged in to the other M1 port. The read results were not affected by the LG in the other port, but the write speeds were faster with LG plugged in.

Difficult to make sense of this lot.

 

[joevt]

Difficult to make sense of this lot.

Not really (except for write speeds being faster with LG plugged in - how much faster?). Different USB controllers have different performance (especially if they are limited by PCIe connection such as the ASM1142 or if they have a problem like the M1 or if they are missing certain features - UASP?)

LG USB: There is a USB 3.0 hub (5 Gbps) connected to a USB 3.x Bus inside the display. Show a screenshot from the USB tab of System Information.app with the USB bus selected so we can see the controller and hub vendor/product IDs. Is it a Fresco Logic FL1100? I forget how the USB is setup in the LG. The LG is connected using Thunderbolt, so the USB must be provided by a PCIe USB controller.

LG TB3: This is a USB 3.1 gen 2 (10 Gbps) capable port provided by the USB controller inside the Thunderbolt controller of the display. > 500 MB/s means the T5's are all connecting properly at 10 Gbps.

M1: Does it show your device is connecting at 5 Gbps or 10 Gbps in the USB tab of System Information.app? Those numbers are worse than 5 Gbps so I suspect the connection is only 5 Gbps. You could try finding a USB 3.1 gen 2 hub that connects at 10 Gbps. Then the Samsung T5 may be able to connect at 10 Gbps while connected to the hub. 5 Gbps from the M1 has lower performance than other USB controllers, so the 10 Gbps connection will probably also be slower, but it least it will be greater than 5 Gbps.

 

[Mike Boreham]

Thanks very much @joevt. That helps a lot. In the order you mentioned:

The increase in write speed for having the LG plugged in to other M1 port was between 20-30MB/s for the three T5s. Not large but real, I did A-B-A tests to confirm.

LG USB. The LG USB ports being 5Gbps is not a issue: the spec is 5Gbps and they give 5Gbps. However you asked for a screenshot, so my USB Device Tree with a 500GB Samsung T5 plugged in to a USB port on the LG Display, is below. I can't see any manufacturer name except LG.

LG TB3. Yes SI confirms 10Gbps so all good

M1. Yes, SI shows the direct connect T5 as connecting at 5Gbps. Why? The M1 MBA specs show USB 10Gbps. I hadn't noticed that SI actually only shows 5Gbps. This is the core issue of the thread. I thought it was claiming 10, but giving 5, but it doesn't even even claim to be 10, except in the specs. I don't have any other 10Gbps drives to test, could it be a Samsung/Apple compatibility issue?

I do have unused OWC TB4 and a Caldigit TS3 Thunderbolt hubs which may give 10Gbps. The lack of 10Gbps is not a daily problem for me as I can keep my T5 plugged in to the LG TB3 port to get 10Gbps. (The recently acquired LG has made the OWC hub redundant, and I had the TS3 for a second Mac I no longer have).

 

[Mike Boreham]

I do have unused OWC TB4 and a Caldigit TS3+ Thunderbolt hubs which may give 10Gbps. The lack of 10Gbps is not a daily problem for me as I can keep my T5 plugged in to the LG TB3 port to get 10Gbps. (The recently acquired LG has made the OWC hub redundant, and I had the TS3 for a second Mac I no longer have).

Out of curiosity I have tested the Samsung T5 using these TB hubs attached to the M1. As expected they both give 10Gbps to the T5 but but the TS3+ is faster than the OWC hub. In fact the TS3+ is a bit faster than the LG.

All with 500Gb Samsung T5, all read/write MB/s :

Direct to M1 MBA (5Gbps): 388/397 (from previous)

LG USB (5Gbps) port: 425/434 (from previous)
LG TB3 port (10Gbps): 522/503 (from previous)

OWC USB (10Gbps port): 446/460
OWC TB4 port (10Gbps): 458/442

TS3+ USB (5Gbps port): 364/342
TS3+ USB (10Gbps port): 529/510
TS3+ TB3 port (10Gbps): 527/510

So some manufacturer variability that you mentioned there.

Incidentally I also tested a Thunderbolt Envoy Express drive on all four Thunderbolt ports available to me, and the results were very similar, much tighter than the USB variations.

 

[joevt]

The increase in write speed for having the LG plugged in to other M1 port was between 20-30MB/s for the three T5s. Not large but real, I did A-B-A tests to confirm.

Just weird that the T5 is faster connected to a M1 Thunderbolt port when the LG is connected to the other port than when it is by itself.

LG USB. The LG USB ports being 5Gbps is not a issue: the spec is 5Gbps and they give 5Gbps. However you asked for a screenshot, so my USB Device Tree with a 500GB Samsung T5 plugged in to a USB port on the LG Display, is below. I can't see any manufacturer name except LG.

Ok, looks like there's no separate USB controller in the display. Instead, you have a Titan Ridge JHL7440 which has a Titan Ridge USB controller with a USB port for the downstream Thunderbolt port and a second USB port for a separate USB port that the Titan Ridge JHL7540 doesn't have (and neither does Alpine Ridge). To verify this, it would be useful to see one device connected to a 5 Gbps USB-C port and another USB device connected to the TB port. Or a dump from ioreg since ioreg shows USB ports even if nothing is connected to the USB port.

M1. Yes, SI shows the direct connect T5 as connecting at 5Gbps. Why? The M1 MBA specs show USB 10Gbps. I hadn't noticed that SI actually only shows 5Gbps. This is the core issue of the thread. I thought it was claiming 10, but giving 5, but it doesn't even even claim to be 10, except in the specs. I don't have any other 10Gbps drives to test, could it be a Samsung/Apple compatibility issue?

Yes, I think it's a combination of Apple Silicon and Samsung. Some 10 Gbps devices connect properly and some connect only at 5 Gbps. It's a problem with USB link speed negotiation.

This problem is different than the known problem with the ASMedia ASM1142 where it shows devices connected at 5 Gbps but they are actually connected and working at 10 Gbps.

I do have unused OWC TB4 and a Caldigit TS3 Thunderbolt hubs which may give 10Gbps. The lack of 10Gbps is not a daily problem for me as I can keep my T5 plugged in to the LG TB3 port to get 10Gbps. (The recently acquired LG has made the OWC hub redundant, and I had the TS3 for a second Mac I no longer have).

The OWC TB4 is good for the M1 Mac because the OWC only works 100% with Big Sur and M1 Macs can't use anything earlier than Big Sur. The OWC gives you four 10 Gbps ports (but the combined USB only bandwidth is only 10 Gbps since it uses a single USB hub for all 4 ports - more USB bandwidth than 10 Gbps requires adding another Thunderbolt device - such as the display).

Out of curiosity I have tested the Samsung T5 using these TB hubs attached to the M1. As expected they both give 10Gbps to the T5 but but the TS3+ is faster than the OWC hub. In fact the TS3+ is a bit faster than the LG.

All with 500Gb Samsung T5, all read/write MB/s :

Direct to M1 MBA (5Gbps): 388/397 (from previous)

LG USB (5Gbps) port: 425/434 (from previous)
LG TB3 port (10Gbps): 522/503 (from previous)

OWC USB (10Gbps port): 446/460
OWC TB4 port (10Gbps): 458/442

TS3+ USB (5Gbps port): 364/342
TS3+ USB (10Gbps port): 529/510
TS3+ TB3 port (10Gbps): 527/510

So some manufacturer variability that you mentioned there.

I think all those new numbers make sense.

I expect all the OWC ports to be the same (USB-C / USB-A) because they are all connected to the same USB hub internally. I expect them to be slower than the LG and the TS3+ because the OWC uses a hub.

The TS3+ doesn't use any hubs. It has two FL1100 for 5 Gbps ports and an ASM1142 for the 10 Gbps port and an Alpine Ridge USB controller for the TB3 port. The ASM1142 is limited to 8 Gbps because it has only a PCIe 3.0 x1 connection but the T5 is slower than 8 Gbps so it doesn't matter.
https://forums.macrumors.com/thread...able-thunderbolt-3-dock.2286171/post-29639488
https://forums.macrumors.com/thread...able-thunderbolt-3-dock.2286171/post-29644981
Did you test the type A port next to the 10 Gbps type C port? I suspect it may support 10 Gbps connection also since it is also connected to the ASM1142. Check the USB tab in SI to be sure it's connected to the ASMedia, and check the benchmark to see if it is > 500 MB/s (since the ASM1142 may report 5 Gbps when it's actually working at 10 Gbps).
https://www.caldigit.com/ts3-plus-interface-bandwidth-allocation-and-diagram/

Incidentally I also tested a Thunderbolt Envoy Express drive on all four Thunderbolt ports available to me, and the results were very similar, much tighter than the USB variations.

That makes sense. The Thunderbolt Envoy is PCIe instead of USB. The only thing that would change its performance is the number of Thunderbolt devices between it and the M1 Mac. I suspect you saw a slight drop when using a Thunderbolt dock or Thunderbolt hub. The M1 Mac uses USB4 so it's limited to a depth of 5 (I haven't seen this verified). A Mac that uses Thunderbolt can chain 6 devices (one more than USB4).

 

[Mike Boreham]

Did you test the type A port next to the 10 Gbps type C port? I suspect it may support 10 Gbps connection also since it is also connected to the ASM1142. Check the USB tab in SI to be sure it's connected to the ASMedia, and check the benchmark to see if it is > 500 MB/s (since the ASM1142 may report 5 Gbps when it's actually working at 10 Gbps).

The Sam T5 the following connected to the various USB ports on the TS3+:

Rear USB-C 10Gbps port: 527/519
Rear USB-A 5Gbps port: 431/415
Front USB-C 5Gbps port: 362/347
Front USB-A 5Gbps port: 362/347

As you expected SI reported all these as 5Gbps connection, even the first which was working at 10Gbps.

The rear USB-A 5Gbps port is faster than the front two 5Gbps, but not as fast as the rear USB-C 10Gbps port.

Thanks very much for your deep input in all this. Now I have read (nearly) all the rest of the thread I realise most of what I have reported is old news and fits the general pattern.

 

[Mike Boreham]

Just weird that the T5 is faster connected to a M1 Thunderbolt port when the LG is connected to the other port than when it is by itself.

Your surprise at this had me worried, so I repeated this again, A-B-A-B

1. 387/396 with LG plugged in other port
2. 385/344 without....
3. 388/387 with......
4. 386/343 without.....

All with Sam T5 connected direct to M1 MBA, R/W, MB/s

 

[joevt]

The Sam T5 the following connected to the various USB ports on the TS3+:

Rear USB-C 10Gbps port: 527/519
Rear USB-A 5Gbps port: 431/415
Front USB-C 5Gbps port: 362/347
Front USB-A 5Gbps port: 362/347

As you expected SI reported all these as 5Gbps connection, even the first which was working at 10Gbps.

The rear USB-A 5Gbps port is faster than the front two 5Gbps, but not as fast as the rear USB-C 10Gbps port.

Ok, those numbers make sense except I don't understand why CalDigit didn't enable 10 Gbps from the USB-A port of the ASM1142 or how it's possible to not be 10 Gbps. I wonder if the ASM1142 firmware could be updated to enable 10 Gbps for the USB-A port (there are ASM1142 firmware updaters for Windows). That would only work if the 5 Gbps is not selected externally.

If the USB-A port is really 5 Gbps from the ASM1142, then it appears that the ASM1142 can do USB 5 Gbps better than the FL1100 can. That's probably because the ASM1142 is PCIe 3.0 x1 (8 Gbps). The FL1100 is PCie 2.0 x1 (4 Gbps) which matches USB 3.0 (4 Gbps) (both use 8b/10b encoding) but maybe PCIe has more overhead than USB or the difference is just what happens when you have two different pipes connected together even if they are the same size - there's a slight misalignment of the pipes that reduces flow.

Your surprise at this had me worried, so I repeated this again, A-B-A-B

1. 387/396 with LG plugged in other port
2. 385/344 without....
3. 388/387 with......
4. 386/343 without.....

All with Sam T5 connected direct to M1 MBA, R/W, MB/s

Interesting. Not sure what it means though. 10% difference is not insignificant. Is it saving power or cycles in the case where only one port is used, which lowers the single port performance? the LG uses Thunderbolt. Do other Thunderbolt devices have the same affect? What if the LG were replaced with a USB device? What if the ports are reversed?

 

[Mike Boreham]

Interesting. Not sure what it means though. 10% difference is not insignificant. Is it saving power or cycles in the case where only one port is used, which lowers the single port performance? the LG uses Thunderbolt. Do other Thunderbolt devices have the same affect? What if the LG were replaced with a USB device? What if the ports are reversed?

It seems to be only the LG in the other port which has the effect of increasing the write speed. Reversing ports no effect.

To add to the list in my previous post:

5. 386/396 with LG plugged in to other M1 port, but ports reversed
6. 383/341 another Sam T5 plugged into the other M1 port
7. 384/352 with a TB3 Envoy Express plugged into the other M1 port

 

[Bob418]

Just bought a WD SN750 1TB drive and put it into a TB3 enclosure. The speed is impressive on my M1 Pro.

 

[Mike Boreham]

Just bought a WD SN750 1TB drive and put it into a TB3 enclosure.

Which TB3 enclosure? I have same blade in a TB3 Envoy Express, which I know is not fast (20Gbps I think). Thanks.

 

[moonwalk]

Just bought a WD SN750 1TB drive and put it into a TB3 enclosure. The speed is impressive on my M1 Pro.
View attachment 1752640

I have a WD SN750 4 TB drive in an Acasis "USB4.0 NXMe M2" enclosure on my M1 mac mini. It tops out at about 1250 MB/s, if I'm lucky. Mostly it gets around 1,000 MB/s, though often it's only pegging in the high 800s MB/s. WD SN750s above 2TB sizes for some reason have slow write speeds. The drive also seems to write faster if some other drive is writing. Go figure. I'm waiting to receive a Caldigit Element hub and I hope this increases write speeds. I'd complain, but at 1000 MB/s it's orders of magnitude faster than my older SSDs (Samsung T5 and WD Blue). I'm also hoping Apple's os updates straighten out these mysteries.

 

[Bob418]

Which TB3 enclosure? I have same blade in a TB3 Envoy Express, which I know is not fast (20Gbps I think). Thanks.

I actually bought a Nekteck 480GB Thunderbolt 3 SSD NVME Hard Drive, and replaced the Internal SSD with SN750. This enclosure has good build quality and runs cool.

 

[InBruges]

Has anyone found a solution to getting SSD's such as the Samsung T7 (which are benchmarking around 800 Read / Write on the Intel MacBook Pros), running at full speed on M1 Macs? [Without having to carry around a heavy Thunderbolt dock)

My results here:

T7 connected to directly Intel 16-Inch MacBook Pro ~ 894MB/s Read and 840MB/s Write
T7 connected directly to Intel M1 MacBook Air ~ 660MB/s Read and 680MB/s Write, but sometimes shows ~400MB/s

T7 is formatted APFS, running the latest firmware and connected using a high quality short thunderbolt cable. I've also tried other USB C cables, but it doesn't affect speed.




PSSD T7:





Product ID: 0x4001


Vendor ID: 0x04e8 (Samsung Electronics Co., Ltd.)


Version: 1.00


Serial Number: S5T4NJ0N802507X


Speed: Up to 10 Gb/s


Manufacturer: Samsung


Location ID: 0x00200000 / 1


Current Available (mA): 900


Current Required (mA): 896


Extra Operating Current (mA): 0


Media:


PSSD T7:


Capacity: 1 TB (1,000,204,886,016 bytes)


Removable Media: No


BSD Name: disk4


Logical Unit: 0


Partition Map Type: GPT (GUID Partition Table)


SMART status: Verified


USB Interface: 0


Volumes:


EFI:


Capacity: 209.7 MB (209,715,200 bytes)


File System: MS-DOS FAT32


BSD Name: disk4s1


Content: EFI


Volume UUID: 0E239BC6-F960-3107-89CF-1C97F78BB46B


disk4s2:


Capacity: 999.86 GB (999,860,912,128 bytes)


BSD Name: disk4s2


Content: Apple_APFS

 

[Screensaver]

Yep. Maybe 11.2 will help? You probably saw, I doubled my read speeds when using the Samsung 970 EVO Plus. But both the Samsung and the ADATA performed essentially the same when put in a normal PCIe NVMe slot on a motherboard. So it's a problem with either macOS, or the enclosure.

Hello there. I saw that you have the ACASIS enclosure and I was wondering if you have since seen speed gains through the misc. Big Sur updates. Have your speeds improved?
Also, have you attempted to install Big Sur on your ACASIS drive and boot from it? Before I buy, I was hoping to confirm if ACASIS you can externally boot from an M1 from it, as well as enable FileVault. I know that some drives are finicky.

 

[Screensaver]

No, just read would be a bug in macOS. Check it out here. Not sure it applies to the M1's though. I will test when I receive my Acasis enclosure.

Hello there! I am wondering, have you since gotten a chance to run #’s on a 2 TB WD Black drive with the ACASIS enclosure? I’m super interested to find out if it cuts speeds in half, like some folks are suggesting.

also, by chance, have you tested whether or not the M1 Mac can boot externally from the ACASIS enclosure w/ WD black? Would love to know if FileVault encryption works too with external boot too.

 

[rJonze]

Hello there! I am wondering, have you since gotten a chance to run #’s on a 2 TB WD Black drive with the ACASIS enclosure? I’m super interested to find out if it cuts speeds in half, like some folks are suggesting.

also, by chance, have you tested whether or not the M1 Mac can boot externally from the ACASIS enclosure w/ WD black? Would love to know if FileVault encryption works too with external boot too.

I used a 2TB pioneer (Phison e12) in the acacias USB4/TB case. Took 10 minutes to install Big Sur and 20 more to MA my user over from the internal on the MBA. Install was straightforward from the recovery screen.Booting from the external works as expected. Read speed is the same as the internal (~2700MB) but write is slightly slower at around ~1700MB vs ~2100 on the internal. Don't use file vault so I can't speak to that. Some have suggested that system software updates don't work properly on external drives but I haven't tried it yet as I installed 11.2.3 from the get go. If the external is not plugged in it happily boots from the internal without bitching. Blade in the external enclosure runs about 5-10º F cooler than the internal as reported by istat menus. User experience seems exactly the same. Having to hang an external box off the air is sub optimal in terms of portability but I can just run without it if I need to. I bought the base model as I was interested in trying out the M1 without making a deep investment. If I was doing it over I'd opt for the 16/512 which seems more like the sweet spot for price/performance.

 

[Peepo]

Has anyone found a solution to getting SSD's such as the Samsung T7 (which are benchmarking around 800 Read / Write on the Intel MacBook Pros), running at full speed on M1 Macs? [Without having to carry around a heavy Thunderbolt dock)

My results here:

T7 connected to directly Intel 16-Inch MacBook Pro ~ 894MB/s Read and 840MB/s Write
T7 connected directly to Intel M1 MacBook Air ~ 660MB/s Read and 680MB/s Write, but sometimes shows ~400MB/s

T7 is formatted APFS, running the latest firmware and connected using a high quality short thunderbolt cable. I've also tried other USB C cables, but it doesn't affect speed.

No, there is no solution. This thread is becoming more confusing due to people discussing thunderbolt enclosures, but the topic of this thread indicates an issue with usb c external SSD being slow like your T7.

I have both a T5 512GB and a SanDisk 2TB Extreme portable SSD and performance is not as expected - while plugged into an M1 Mac it seems to be only using about 70%-80% of speed compared to using an Intel Mac, or Windows PC.

 

[Fomalhaut]

I have a related question: which 3.1 Gen 2 to SATA-III adapters are recognized as supporting a 10Gbps link?

I'm talking about cables like this one: https://www.amazon.com.au/Plugable-...USB+3.1+to+SATA+adapter&qid=1617926152&sr=8-2

...which connect a naked 2.5" SATA drive to a USB 3.1 gen 2 connector (either USB-C or USB-A)

Currently, the connection speeds with USB 3.1 gen 2 (or 3.2 gen 2x1) devices seem hit and miss on the M1 Macs. I already have a 5Gbps adapter cable, and don't want to waste money on a new 3.1 gen 2 cable if it won't connect at 10Gbps.

[Edit: looks like many of these adapters use the AS Media chipset. Does the chipset family consistently *fail* to connect at 10Gbps on the M1 Macs or only some variants? ]

 

[legato01]

I have a related question: which 3.1 Gen 2 to SATA-III adapters are recognized as supporting a 10Gbps link?

I'm talking about cables like this one: https://www.amazon.com.au/Plugable-...USB+3.1+to+SATA+adapter&qid=1617926152&sr=8-2

...which connect a naked 2.5" SATA drive to a USB 3.1 gen 2 connector (either USB-C or USB-A)

Currently, the connection speeds with USB 3.1 gen 2 (or 3.2 gen 2x1) devices seem hit and miss on the M1 Macs. I already have a 5Gbps adapter cable, and don't want to waste money on a new 3.1 gen 2 cable if it won't connect at 10Gbps.

[Edit: looks like many of these adapters use the AS Media chipset. Does the chipset family consistently *fail* to connect at 10Gbps on the M1 Macs or only some variants? ]

If you want to hit 10Gbps or more speeds (in theory anyway) you'll need NVMe SSDs.

For older hard drive style SSDs, limiting factor is SATA. You won't hit 10Gbps because SATA III maximum throughput is 6Gbps.

 

[joevt]

If you want to hit 10Gbps or more speeds (in theory anyway) you'll need NVMe SSDs.

For older hard drive style SSDs, limiting factor is SATA. You won't hit 10Gbps because SATA III maximum throughput is 6Gbps.

There are hardware RAID 10 Gbps USB enclosures which can get near 10 Gbps with two SATA III SSDs.
https://eshop.macsales.com/shop/external-drives/owc-mercury-elite-pro-dual/mini
but I would stay away from RAID.

 

[Tenkaykev]

I've been following the thread and I'm about to purchase an external SSD for backups. I've decided on 1TB and the two contenders are Sandisk Extreme portable and the Samsung T7.
They are both the same price and looking at the spec it seems that they are both capable of the same transfer speeds ( I appreciate that my M1 Air won't achieve the speeds quoted )
Does anyone have any real world experience with these drives which may help my purchase decision? TIA

 

[Peepo]

I've been following the thread and I'm about to purchase an external SSD for backups. I've decided on 1TB and the two contenders are Sandisk Extreme portable and the Samsung T7.
They are both the same price and looking at the spec it seems that they are both capable of the same transfer speeds ( I appreciate that my M1 Air won't achieve the speeds quoted )
Does anyone have any real world experience with these drives which may help my purchase decision? TIA

Samsung T7 Touch Portable SSD Review: Sleek, Secure Pocketable Storage

Samsung's T7 Touch portable SSD lets you unlock your data with a swipe of your finger.

www.tomshardware.com

Definitely the Sandisk.