Testing Mac Mini M2 USB-A port speed help needed

[vertumno]

Hello,
I'm testing the USB-A ports of my new Mac Mini M2 (octacore) and I found some unexpected results.

In short if I connect just an SSD to, let's say port 1, I get the speed expected for that particular drive, BUT if I connect also a usb hub (I've tried 3 all powered) to the other USB A (let's call it port 2) the SSD on port 1 read speed decreases from 400 MB/s to 150 MB/s.

It's like the ports are interfering with each other/sharing the same bus (?)

Tried the same configuration on an old Mac Mini i7 and no decrease in speed, at the Apple store they don't know and wouldn't try on their machines.

Is anyone here willing to take the same test and see if it's a design choice or if I have a problem on mine?

The test would be:

- connect and SSD on USBA port1 and nothing else on port2, measure the speed with Blackmagic or similar

- connect the same SSD on the same port1, but adding a usb hub on port2, and measure the SSD speed on port1 with Blackmagic

Thanks in advance to anyone willing to try this out

 

[Bigwaff]

Some more specifics about your USB hubs and USB throughput they support would be helpful, including whether you are attaching devices to them when you run your tests.

Long running thread about USB speeds on M-series Macs.. mostly related to M1 series but later discusses M2 series have the same USB speeds. Someone actually did some testing similar to what your are requesting ....

https://forums.macrumors.com/threads/usb-on-m1-macs-isnt-actually-10gb-s-also-definitely-not-usb4.2269777/post-31762971

 

[vertumno]

Thank you but the problem most definitely doesn’t lie in the hubs per se, as you can read on a old Mac mini i7 speeds we’re actually better. So either mine is toasted or the usb bus architecture has changed for the worse.

I can really only solve this riddle if someone with a similar machine is so generous to waste some time in replicating the problem.

 

[Bigwaff]

I can really only solve this riddle if someone with a similar machine is so generous to waste some time in replicating the problem.

If you take the time to read the thread I linked, you will discover the issue is related to interactions between third-party USB controllers, Apple's internal USB controller and Apple's USB driver implementation on Apple Silicon. Intel Macs use different internal USB controllers than Apple Silicon Macs.

 

[vertumno]

Thank you I’ve read it but is a different model and I get very different readings, especially the fact the read speed is a lot slower than write speed is worring, could not see anything that extreme on the tests in the thread provided. It might well be what you say, but it will be nice to have a more direct comparison between similar settings.

 

[meson]

I run an M1 mini 16/256.

I just connected my WD Blue SSD in an SSK enclosure to one usb-a port. I get about 300 MB/s read/write. I connected a simple usb3 hub (nothing else attached) to the other usb-a port. There was no appreciable speed change with the ssd. Tested using Blackmagic Disk Speed Test.

That same drive runs at about 500 MB/s read/write when connected to a 10 Gb/s usb-a port on my Monoprice Thunderbolt hub, and maybe about 650 MB/s when connected directly via usb c. There is an overhead penalty associated with daisy chaining devices with usb controllers. Are you connecting your ssd to a hub? Do you have other peripherals plugged into the hub? Or are your speeds affected merely by connecting the bare hub to the second usb-a port?

I will say having both usb-a ports occupied destabilized the Bluetooth connection with my Magic Trackpad (another known and enjoyable issue with the mini), even with the mini on its side and edge with the ports and 'bottom' of the machine having a direct line of sight path to my peripherals.

 

[Fishrrman]

OP:

Do you have a DIFFERENT USB3 hub that you could try?