3 ms·

Anyone that has hooked up a external drive to an M1 or M2 based notebook knows why: the performance of external devices dropped off a cliff with Apple’s own sil

Anyone that has hooked up a external drive to an M1 or M2 based notebook knows why: the performance of external devices dropped off a cliff with Apple’s own silicon. Not to mention single external support… Consumers Pro users are likely to look aside on the external IO issues, but it’s doubtful that actual studios would.

sephamorr 4y ago
Can you tell me more about this? Are you unable to hit line rates, or are you just commenting about lackluster io options? I'm looking to add some M.2 storage via thunderbolt enclosure to my machine and would otherwise be expecting to be able to saturate the pcie lanes.
- exabrial 4y ago
  This regression I've personally tested and confirm: You can't hit anywhere near "line rate" with usb 3.1/3.2, though an Intel Mac will gladly do so with the same _exact_ drive. You will experience write stalls and other performance problems. I don't have a usb4 drive to test by I wouldn't hold your breath. I do not have a Thunderbolt 3 device to test. Google tells me it's a problem: https://9to5mac.com/2022/04/18/m1-mac-thunderbolt-4-ports-speed-tests/ https://9to5mac.com/2022/04/18/m1-mac-thunderbolt-4-ports-sp... Expect _at most_half the advertised speeds and feel lucky if you achieve them. EDIT: From the article I linked: I just noticed they suggest try connecting my drive to a Thunderbolt 3 dock first. LOL, ok, I'll give this a try.
CharlesW 4y ago
> Anyone that has hooked up a external drive to an M1 or M2 based notebook knows why: the performance of external devices dropped off a cliff with Apple’s own silicon. This is false in my experience. I bought a Thunderbolt 3 to Dual NVMe M.2 SSD enclosure (Sabrent EC-T3DN), set up the SSDs in a RAID 0 array, and Blackmagic Disk Speed Test benchmarked that configuration at 2,500 MB/sec read and write. That's very close to the maximum real-world usable bandwidth for TB3¹. ¹ https://macperformanceguide.com/blog/2019/20190128_1352-understanding-Thunderbolt3-bandwidth.html https://macperformanceguide.com/blog/2019/20190128_1352-unde...

sephamorr 4y ago

Can you tell me more about this? Are you unable to hit line rates, or are you just commenting about lackluster io options? I'm looking to add some M.2 storage via thunderbolt enclosure to my machine and would otherwise be expecting to be able to saturate the pcie lanes.

exabrial 4y ago
This regression I've personally tested and confirm: You can't hit anywhere near "line rate" with usb 3.1/3.2, though an Intel Mac will gladly do so with the same _exact_ drive. You will experience write stalls and other performance problems. I don't have a usb4 drive to test by I wouldn't hold your breath. I do not have a Thunderbolt 3 device to test. Google tells me it's a problem: https://9to5mac.com/2022/04/18/m1-mac-thunderbolt-4-ports-speed-tests/ https://9to5mac.com/2022/04/18/m1-mac-thunderbolt-4-ports-sp... Expect _at most_half the advertised speeds and feel lucky if you achieve them. EDIT: From the article I linked: I just noticed they suggest try connecting my drive to a Thunderbolt 3 dock first. LOL, ok, I'll give this a try.

CharlesW 4y ago

> Anyone that has hooked up a external drive to an M1 or M2 based notebook knows why: the performance of external devices dropped off a cliff with Apple’s own silicon. This is false in my experience. I bought a Thunderbolt 3 to Dual NVMe M.2 SSD enclosure (Sabrent EC-T3DN), set up the SSDs in a RAID 0 array, and Blackmagic Disk Speed Test benchmarked that configuration at 2,500 MB/sec read and write. That's very close to the maximum real-world usable bandwidth for TB3¹. ¹ https://macperformanceguide.com/blog/2019/20190128_1352-understanding-Thunderbolt3-bandwidth.html https://macperformanceguide.com/blog/2019/20190128_1352-unde...