How to stress test I/O on usb3 hard drive?

[Ecchi Nutuano]

I have a usb3 device, that intermittently disconnects and immediately reconnects(with the usual warning this device was not ejected properly blah blah blah). The only correlation I think, is that it *sometimes* happens on heavy I/O. More like massive amount of small file transfers than big sized ones. I'm trying to figure out if its a problem with the usb3 hard drive, or hackintosh. It doesn't happen all the time either, which makes troubleshooting it that more difficult. Power stability should not be an issue, runs on UPS.

This is the only usb3 device I possess.

I can test it against a current Macbook Air. But I would like the tests to be identical to eliminate the possibility that the device is defective or not.

I'm leaning to hackintosh as being the problem. The USB3 on my GA-Z68X-UD3H-B3 (rev. 1.0 - running EFI U1H) from what I can tell is 3rd party chipset. It was before intel made their own supported USB3 in its chipsets. Which may explain the occasional instability. Windows has specific drivers for it from Etron.

Stress test suggestions are welcome.

I'm running 10.9.4, but I've witnessed this behaviour through all the 10.9's, and I think 10.8's (the window I was running 10.8 before moving to 10.9 when I bought this hard drive, is so small, I can't say with certainty).

 

[AppleIIGuy]

Blackmagic Disk Speed Test?
https://itunes.apple.com/us/app/blackmagic-disk-speed-test/id425264550?mt=12

 

[Ecchi Nutuano]

Tried it. Ran it for awhile with no issue. Examples of when this behaviour seems to kick in:
back-ups to the drive (tonnes of tiny little files)*
un-rar/taring/ziping huge archives onto the drive*.

* = doesn't happen all the time. On some disconnects of these actions, the very next attempt it won't happen again

Guess its the problem with an intermittent issue. I'm trying to reliably create a reproducible disconnect, then perform the same action on a legit mac. About the only way to know for sure if its my mobo's 3rd party USB3, or the USB3 device that is at fault.

 

[WonkeyDonkey]

Tried it. Ran it for awhile with no issue. Examples of when this behaviour seems to kick in:
back-ups to the drive (tonnes of tiny little files)*
un-rar/taring/ziping huge archives onto the drive*.

* = doesn't happen all the time. On some disconnects of these actions, the very next attempt it won't happen again

Guess its the problem with an intermittent issue. I'm trying to reliably create a reproducible disconnect, then perform the same action on a legit mac. About the only way to know for sure if its my mobo's 3rd party USB3, or the USB3 device that is at fault.

Have you tried AJA System Test ?

Link.

 

[Ecchi Nutuano]

Thanks. In "Sweeping" mode, it effectively causes the disconnect on my hackintosh. Very reproducible (though duration varies).

Tried it on the MBA, and it ran rock solid.

Tried it on a MP1,1 which is running a 3rd party startech USB3 PCIe card (used a hacked/patched caldigit kext IIRC). Could access content, but AJA kept stopping immediately on start. Tried plugging the card into my hackintosh, and although system profiler could see new usb3 super speed, drive wasn't mounting with it. Weird. Could be how the patched caldigit kext identified it or whatever. Not going to mess with it further.

So the problem is clearly with the built-in USB3 on my motherboard. A shame.

 

[WonkeyDonkey]

Thanks. In "Sweeping" mode, it effectively causes the disconnect on my hackintosh. Very reproducible (though duration varies).

Tried it on the MBA, and it ran rock solid.

Tried it on a MP1,1 which is running a 3rd party startech USB3 PCIe card (used a hacked/patched caldigit kext IIRC). Could access content, but AJA kept stopping immediately on start. Tried plugging the card into my hackintosh, and although system profiler could see new usb3 super speed, drive wasn't mounting with it. Weird. Could be how the patched caldigit kext identified it or whatever. Not going to mess with it further.

So the problem is clearly with the built-in USB3 on my motherboard. A shame.

I ran the matx version of the motherboard you have and never got USB3 working either. It was some time since I tried it though.