[LAU] Analysing (debugging) USB port performance

[Jeremy Jongepier]

Hello Lorenzo,

> I have anecdotally, but recurrently observed on laptops - including a
> rather recent one - that there is always a 'best' USB port for external
> sound cards.
>
> For instance, on my latest machine with a decent 'realtime audio'
> configuration/set-up (real-time kernel, /etc/limits stuff, 'performance'
> CPU governor, (wireless) network switched off), I'm able to have a
> pleasant xrun-free session recording in Ardour including a bunch of
> tracks with effects playing at 64 frames and period of 3 with a
> relatively cheap card (UMC202) on one of the USB ports.
> On the other hand, in the exact same conditions I get incidental xruns
> at even 128 frames and xrun instability at 64 frames on the other USB
> ports.
>
> I wonder:
>
> 1. Is there a more scientific (well, precise at least) method to assess
> this USB port performance? What to test or look into?

You can check with lsusb what sits on which bus and if your soundcard
sits on the same bus with a device that could interfere (Wifi,
Bluetooth) then either try moving it to a different bus or try freeing
that bus. A bus is not the same as an USB port, multiple USB ports could
be hooked to the same bus.

> 2. Is there a way to change (e.g. improve the not-so-good USB port
> performance) OS/software wise, or is this usually hard-wired in notebooks?

You can prevent kernel modules from loading by blacklisting them or even
unbinding them:
https://autostatic.com/2013/12/29/resolved-jack-issues-on-notebook/

>   2a. Are IRQs relevant on laptops and if so can a whole USB port (or
> the device attached to it) be optimised from the OS?

Yes they are and they can be optimised if you use threaded IRQ's (i.e.
hardware IRQ's that have a software counterpart). There is useful info
about that on https://wiki.linuxaudio.org/wiki/system_configuration

> Of course I _can_ live with one 'good sound-card port' on a laptop but
> I'm quite curious about people's experiences and the gurus' wisdom -
> albeit on my former machine this was the left-side port which was closer
> to where the sound-card usually sits, now it's on the right, too bad!
>
> Hopefully other LAU have mused about such USB-related mysteries in the
> past...

Well I have so if you'd like to improve your current situation I might
be able to help out.

Jeremy



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: OpenPGP digital signature
URL: <https://lists.linuxaudio.org/archives/linux-audio-user/attachments/20190925/2b7a33d0/attachment.sig>