On Sat, 14 Jun 2014, Ralf Mardorf wrote:
If you add all users to the audio group, then some users might experience issues, when they don't need RT audio, but they e.g. want to use transparency windows with FLOSS 3D support, when a full preempted (lowlatency) or kernel-rt is installed. Perhaps such a kernel already does cause issues, even if the user isn't in the group audio. However, my user always join this group and the RT vanilla kernels provide doesn't drastically slow down GUI performance when using transparency, but preempted and preempted rt kernels can not be used.
As a user of the lowlatency kernel as my default boot (in fact the only kernel installed) I don't have any graphics issues from this or being a member of the audio group.
JFTR, keep in mind that for RT nice values are nonsense.
Yup, that is why it is commented out.
Likely memlock could become an issue too.
I have had no issues... unless memory is too small (< 1G) but then there will likely be problems for many graphics creation applications anyway. In any case the problems would only show up if there is an application running that uses memlock. This would be one of the best arguements against running jackd(bus) as session audio backend.
And please search LAU, LAD, and I guess especially jack-devel threads for frantic efforts to provide apps that automatically do set up jackd latency settings. IT'S IMPOSSIBLE!
Set up latency? no but we would like to automatically find out latency limits. The upper limit doesn't seem to matter too much jackd just silently sets the highest it can. For example setting my ICE card for -p4096 has the same effect as setting it to -p2048. However, setting jack to run at too small a latency setting (for example setting an Intel HDA to 64/2) will crash jackd. I would like to avoid having a zombie jackd running that the user has to kill somehow before they can restart jack.
This is not about the lowest "stable" latency setting for a card. My ICE card can be set and start with -p16, but only run at that latency in a stable manner with hyperthreading turned off and up to certain CPU use levels. The user will have to learn what the lowest latency is that they can use for a certain task without xruns. We can't do that. I would like to have an "About Latency" button for this reason.
Don't write an app that fakes to support something, that cannot be supported.
No worries. -- Len Ovens www.ovenwerks.net -- ubuntu-studio-devel mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-studio-devel
