Has Apple FIXED USB Speeds and Bluetooth with the NEW M1 Pro & M1 Max chips?

50 Comments

1. If you format drive to APFS format then you will increse speed of ssd drive

2. 2023 and still USB issues. Samsung T7 on a 2018 MBP read and white around 890, on a 2022 M1, read and write around 720. It's insane. I would have never gotten one if I knew this. I've would have bought the 2019 MBP instead. Yes, I consider having better read and write speeds more important.

3. Has fixed with Ventura 13.2 ?

4. Almost a year later and this is still an issue! Apple clearly cant or won't fix the problem. Let's hope the M2 chips resolve it.

5. and of course an update. Its a travesty that ppl have to wade through 1000s of billions of terabytes of info on the net, taking forever even with dedicated smart searching, to find info…. but here it is after my earlier post… from dpreview, one of my fave AV sites…

   "The dual-lane (2×2) mode gets you "up to 20 Gbit/s", but there's a catch. For USB 4 host ports, support of Tunneled USB 3.2 (20 Gbit/s) is optional. I'm a little bit surprised that if Apple went to the trouble of including support for 40 Gbps Thunderbolt 3 and 40 Gbps USB4 transfer modes, that they left out support for 20 Gbps USB 3.2 2×2 mode, but the USB4 standard does allow that; and there you have it."

   and may i add the obvious, left out support for anything beyond 3.1 gen2x1

   Apple didnt 'make a mistake', they deliberately chose each element that is in every controller and chip and port their product has. Deliberate. As in THEY KNOW FULL WELL how things 'should be', and chose to cheap out, cop out, dodge out of the best they can do to cut corners and grab more cash to performance ratio.

   Then they marketed as supporting USB4 and its speed… which even the studied wrongly thought will support the 3.x standards below it… while they BARELY support USB at ALL. Hell I coulda saved huge from buying a usb-c connected anything, and gone for something maxing the USB A 5Gbps common port…

   This isnt a mistake this is deliberate underwhelming design. I cant even begin to wonder how many millions of overspending clueless or sometimes even studied customers have committed to, thinking their extra money would gain actual extra performance under the specs.

   So now one has to decide: Since they KNOWINGLY failed me in this area, do I dare pay TWICE AS MUCH AGAIN for an alleged TB3 or 4 device and expect to be disappointed AGAIN?

   You screw with my trust, you wont live it down. Apple did that to themselves. AGAIN. This time instead of reading about everyone else getting caught, Im in the same boat as those on all the forums. FFS Apple… MAYBE you should stop being the 'pioneer of pewp emojis' for once, and focus on WHAT YOU USED TO BE good at making.

6. I have an Nvme m.2 rated 3500MB/s in an external case that is USB3.2 gen2x2 (should get 2384 MB/s) …. I DO NOT get anywhere close to the 2400MB/s that should max out at, I get barely 1/3. Thats on my new M1 Mini I bought last week. I dont even crack 900MB/s.

   I will double check by taking back the external combo and having them test it in the store, but at this point its clear Apple is pumping out poor USB bus stats and hiding behind hyped claims… blame Apple controller or hyped advertising, whichever.

   The ads and product spec all say "USB4 (40Gb/s)/Thunderbolt 3 (same)" so you would THINK the port would support USB4 (I BET IT WONT), and backward compatible to whatever the lesser USB standard speed is for whatever protocol plugs into it…

   I check the system report, I see both USB ports are 3.1 ONLY. Ah, now I understand why Apple gives itself an out about USB with the other spec saying 3.1 gen 2×1 when USB-c. Theres no fixing that, or firmware flashing, or system updating with fixes, only retrofitting by 'a newer M1 Mini USB Controller' sometime in the future. Screw that. Im sick of half truths and half promises from apple, decades now.

   Sure enough, connecting the USB3.2 combo it SHOWS UP as 3.2 underneath a 3.1 controller bus, and is assigned 3.1 status of 10Gb/s (1192 MB/s) not 20Gb/s (2384 MB/s).
   Think you solve it with a TB3 powered dock? No..
   plug it into a TB3 dock, check system report, dock shows up under the TB category getting the full 40GB/s, and the USB controller reaches downsrream and assigns the unit 10Gb/s, again. The dock DID NOT help; MORE money spent on something that didnt resolve things, and now Ive spent $200-300 more on extras than I needed to. I couldve just gone for lesser and gotten same performance, but for that APPLE WOULDVE HAD TO BE MORE OPEN AND HONEST about its claims and flaws.

   So unwittingly, I paid twice the money for overkill solutions that Apple WILL NOT CANNOT make even 1/3 use of. Lesson: DO NOT buy up, buy DOWN for Apple accessories from now on. And dont believe the hype, EVER.

   ""So what, you are getting better performance than your previous machine right?""
   THATS NOT THE POINT.
   THE POINT is, via shifty sketchy biz practice and marketing, I spent far more money than needed for things I CANNOT make use of, which I consider a WASTE. Im not a money tree.
   I also DISLIKE IMMENSELY dealing with anyone or any thing that has even a wisp of dishonesty to it… where my purchasing decisions can go so far off the rails based on advice suggestions and what I thought was offered (USB4 doesnt even have a controller in the thing, Apple must mean CONNECTION interface not performance/controller inclusion)

   There are accessory makers milking ppl by reassuring you can plug into TB3 for that 'full 40Gb/s experience' while at the same time using USB-C to provide something less, knowingly. Im not certain now that Apple isnt a part of the same crowd of convoluted truth skewers, doing the same thing, and generating a hollow marketplace.
   Way to go Apple… I didnt get conned by the third parties and their hyped claims about your tech, I got conned by overly trusting you guys, that keeps hiding behind the interface potential not the actual bus and port stats. Not even a 3.2 USB controller… REALLY?? SERIOUSLY?

   THAT I didnt expect and is a massive disappointment.

7. I just tested a 128GB CF Express card by Lexar using the Lexar reader connected to the M1 Max thunderbolt port and got 843MB/s write and 909MB/s read speeds. The card itself is imprinted with 1750MB/s.

8. Wish I'd seen this before dropping $4300… Wow… Just, wow. This is some BOOLSH*T.

   Thank you for the thorough video, sir!

9. man it's May 2022 and the USB Speeds are still slow as hell! it doesn't make any sense. all my samsung and sandisk drives run exponentially slower on an M1 Mac. so dumb.

10. Its CLEARLY a driver issue. Using crystaldisk mark on a Windows ARM virtual machine through parallels, i get 983 read and 944 write on Samsung T7.

    But at the same time, transferring real world, the T7 on macos on a 14 inch macbook pro gets 810MB/s both ways in actual real transferring.

    Seems like this is also partially Blackmagic's fault.

11. Not just external drives, but it actually affecting all USB devices
    With USB to Ethernet Dongles (Realtek Gigabit chipset) on M1 (2020 MBP), I can get only 5xx-8xxMbps (which also depends on the cable/router connected to)
    The same dongle can achieve a stable 93x-94xMbps on a Windows PC
    To get same level of network performance, I need to use a Realtek 2.5GbE and run it in Gigabit to achieve 900+Mbps regardless to what device/cable connecting to

12. Brizzel. Happy New year!. A comment for Senior Algor.R.

13. Have you tried testing external drives by using a stopwatch and copying the same file?
    Just to verify that the issue doesn’t lie in the black magic software.

14. Thank you!!! After watching hours of old (Late 2020, Early 2021) YouTube videos complaining about M1 MBP and Mac mini Thunderbolt 4/USB-C 4 ports and external SSDs or NVMe drives, you have answered the question: Is Apple holding back external drive speeds via Thunderbolt?!? YES IMHO!

    My M1 13" MBP is officially over a year old and is the base model, so I want more external storage for archiving data, but I have to transfer said data to newer drives, then I can start more video editing and graphic work. I have edited very short videos for a local band but would love to do more video editing for the small restaurant I work for, where I do clerical work and graphic designs for the menu. So my current use case with the base model M1 13" is fine…

15. Great review. I've just grabbed an Amazon " Black Friday " deal for the T7 1TB for £84, super pleased and the performance will be more than good enough for my usage.
    EDIT:
    The T7 arrived and I connected it as is via the supplied cables. Using the USB A to USB C via an Anker 5 way hub Read speed 314 Write speed 313 MB/s. Connecting directly to the M1 Air via USB C saw 684 MB/s read/ 615 MB/s write. I reformatted the drive to Mac OS extended / journaled and saw the same performance. There was no option to format to AFPS.

16. Apple needs to fix the thunderbolt transfer speeds. – That’s not acceptable.

17. Is it possible the difference in drive performance is something to do with the way MAC OS buffers IO data? I had issues with M1 MACBOOK AIR using Bluetooth mice where every 5-6 seconds I would get stuttery communication for about 0.3 seconds, it was a regular thing. This went away when I updated to MAC OS MONTEREY, so I think Apple did something to rectify this, prior to Monterey I had to switch my Logitech people mouse to USB mode and use the dongle instead when playing FPS games, but now I can use Bluetooth with no issues.

18. I benchmarked an NVMe Samsung Pro 980 on both a 2016 MBP Intel and 2021 MBP M1 Max. I used a cheap (<$100) but good USB 4.0 aluminium enclosure from a Chinese vendor keeping temperature at constant 49 degrees Celsius under Blackmagic test over 10 min. On the 2016 MBP, the USB 4.0 was faster than the internal SSD (~2400 MB/s). On the 2021 M1 Max, i measured 2800 MB/s for both read and write which is very fast alzhough slower than the internal SSD.

    (Note: The vendor, Acasis warns against some SSD models which aren‘t suitable for this level of load – after you make sense of their translated Chinese 😉 )

    So, any test using USB 3 seems useless to me. Of course, on a dock with USB 3, the perfomance folds to a fifth or so …

    Of course, the USB 4.0 drive works on older USB notebooks to. But on USB 4 notebooks, or a less than 5 years old Mac, it flies 🙂

19. Good insights, thanks. The bit with the buds made me laugh a bit. Loading off your old ones to your wife & keeping the new ones for yourself. I'm sure you meant it well, but one could also think that you may have a habit of giving old gadgets to your wife! :)) That would make a good sketch.

20. This was the video I was waiting on. USB performance! I mostly love my M1 MBA but only single monitor support and poor USB performance with external drives and some peripheries just not working properly through the port really sours the experience.

    I was hoping USB would be fixed with the new pros but it’s not! 🤦‍♂️ what is Apple doing here. I was already on the fence about cost and not having Face ID but the crap USB, that’s the final nail.

21. Great video! It is disappointing that the M1 chips are quite a bit slower with both the T5 and T7 drives. I hope that Apple is able to fix that. I hope that they care because that’s a major issue. I can hardly wait for your next comparison video where you’ll be including your Threadripper Pro workstation in the mix!

22. Why are my comments deleted?

23. Oddly enough, I have that Samsung T5 drive, and on my M1 Mac mini connected directly I get around 372 Write and 388 Read. When plug in my Qwiizlab UH25, I get 447Wr. and 489 Rd. in line with what you showed in a previous video. I bought it on the strength of your recommendation, BTW, and it has been performing really well!

    Only oddity was that I added an NVME drive to its storage bay, and when I was first copying large amounts of data, it would disconnect from time to time, but has behaved perfectly as my sound library storage device for Logic and associated sound libraries. 
    PS the NVME in the Qwiizlab UH25 reaches 730Write and 686 Read.

24. Nobody has done a video of the LOW POWER MODE performance of M1Pro or M1Max mbp14/16.

    If battery life is what people seek, I think it's a good compromise to sacrifice some performance to uber long battery life.

    That's just me.

    My dream setup:

    MBP16 M1Pro 32GB, 16c GPU, 10c CPU, 2TB SSD, Low Power Mode on battery, High Power Mode on cable.
    iPadPro 12.9 w/ApplePencil running SideCar for extended display (and touch ability)
    Anker power eliteIII USB-C 25400maH 45-65W PD powerbank for extended battery life.
    iPhone 13 Mini w/Magsafe Battery.
    2x HomePod Mini
    4xAirTag

    this config is gonna absolutely rock.

25. The low level of signal for the keyboard and mouse is a problem with Bluetooth connectivity.

26. Could this be because the Samsung drivers are x86 and run under rosetta 2?

27. A few YouTubers I watch had Bluetooth issues on their M1 Mac minis and getting replacements fixed their issues, which would certainly imply something hardware-related (at least in part).

28. Thank you ..on point as per usual.

29. Thanks for the video.
    There's an effort to port Linux to the M1/Pro/Max Macs, and it will probably be stable within few months, and then it will be a simple test – how's the transfer in Linux and compare it to the latest Mac OS.

30. I’m curious how something like the Lacie Rugged Pro Thunderbolt 3 SSD performs with different cables.

31. hmmmm confusing. Drivers? Hardware supplier quality? Well done Dave. Although Its pretty confusing.

32. With regards to the TB versus USB speeds: I absolutely suspect the operating system. From my knowledge it tries to identify the drive, if you go into the system information, it will show up. There has to be some optimization failure in the driver / operating system.

33. Yes definitely disappointed with the new M1 chips for external storage. Everyone else goes on about performance yet forget to mention half speed transfer rates. How is one supposed to use an M1 for photography or video when the transfer rates are faulty. Even more complicated by the fact different enclosures are hit or miss with different brands of memory or disc controllers.

34. Thanks for that video.
    Curious to know test result with Samsung X5 that take profit of the Thunderbolt.

35. Great stuff. Thanks as always the clear, descriptive content. I’m using an external SSD enclosure with an NVME drive and am pretty happy with the performance on my M1 Air. I know the T5 is more ubiquitous and therefore makes for an easier comparison, but me interesting to see if other manufacturers’ drives or enclosures are also slower in Apple Silicon.
    I’ve not had an issues with Bluetooth either.
    Oh, I just noticed that you’re using the same Dell monitor – great choice!

    Keep up the great work – looking forward to the video editing comparison. Will you be including the tired Grandad Adobe Premiere in your tests? Very mediocre performance on M1 so far, especially compared to the extremely well optimised Davinci Resolve.

36. Thank you for the latest figures on the new MacBooks’ BT connectivity, Dave. Pity they are not as good as they should be, but I still plan on getting a T7 drive, so thanks too for the link.

37. I was frustrated by the outcomes of some of these tests, as I’m sure you were too, Dave. Come on Apple!!! Really?!?!? They had to know about this issue going back to the M1 SoC, and now it’s pretty much of the same! How can they let this slip by with their new high performance laptops. Totally inexcusable, IMHO. Apple engineers need to work with Intel engineers to get this handled. Is this bad blood residual? Perhaps. It should have been addressed. Well done, sir.

38. Have you tested thunderbolt external ssd?

39. Just a note that the i9 iMac 27 has 10 cores, not 8.

40. Glad you reminded me that I needed another T5. I used your Amazon link, toggled it over to 2TB and ordered one to use for time machine backups to migrate stuff to my new MBP when it gets here next week and then to continue to keep portions of it backed up on a nightly or weekly basis. GREAT timing. THANKS!

41. Any video that talks about Solid States gets my approval

42. We are in a cleft stick. I want my new MacBook Pro 14" M1 Pro to last as long as my previous MacBook, 9 years. But anything fails, the whole thing fails, and an obvious point of failure is the internal SSD.

    So I want to use more external storage to preserve the life of the internal storage. This should be straight forward, but apparently not.

    What are your thoughts on the life of the SSD and what to do about it?

43. Great vid! Why is the vid going warm to cool back to warm… I thought it was dramatic effect to start with

44. Did you happen to test actual file transfer speeds and not just benchmark? Like some really large files and see if it makes any difference between Windows, MacOS and different drives? Just curious if that slower benchmark read/write speeds translate to real world.

45. Thanks for the vid. Please make a logic pro test 😀

46. Perhaps Apple can't use the propper drivers, maybe rights owned by Intel and they had to use make shift drivers.

47. Nice cool shirt.

48. Please use my tip to test Topaz Video AI in your M1 Max, thanks!

49. Bluetooth and WiFi is a Broadcom BCM4378 Wi-Fi 6/Bluetooth (BT) 5.0. Turns out, this Broadcom chip doesn't work well with some Bluetooth devices. Logitech explicitly calls out support for Low Energy.

50. How this reflect in real life? 10GB folder? Did you tested this?