Yes, we are dealing with very low latencies. Our embedded device is a pinball machine. Yes, the performance improvements in user-mode are measurably better just by looking at percentages in top. I can't even get playback in system-mode to be without static (i don't know what else to call it). In user-mode it sounds perfect.
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐ On Friday, July 24, 2020 9:27 AM, Arun Raghavan <[email protected]> wrote: > On Fri, 24 Jul 2020, at 8:29 AM, jtharkey wrote: > > > We're trying to get away from running PA in system mode on an embedded > > device because of performance. We have the problem that our app must > > run as root, however. Right now I'm doing this: > > We have two users on the system, root and the app's > > The app's user logs in and runs PA > > The app's user sudos the app to run as root > > This works great except it breaks bluetooth audio support. Is this the > > correct way to go about this or do we need to change our approach? > > Are you working with very low latencies, or such? And have you been able to > measure a performance improvement in user-mode? > > Because outside of low latency use-cases, I'd expect the memcpy() overhead on > most modern systems is small enough to not be a deciding factor. > > Cheers, > Arun _______________________________________________ pulseaudio-discuss mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
