Re: [fluid-dev] OOM killer synths and ALSA overruns

[Josh Green]

The ALSA sequencer buffer overrun error is a bug in FluidSynth, which
really warrants a release, in and of itself.  Its fixed in CVS.  As for
good low-latency Linux response, not sure what the best options are
these days, since I haven't played with it much with newer kernels.  I
have heard though that unpatched the 2.6.x kernels can be either very
bad or moderately OK as far as latency, depending on which particular
version, etc.  I think there are existing tools which help one to track
down sources of latency, since a single driver could in fact be the
culprit.  Much luck.  And if you do figure out a good combo and feel
like sharing the info, please do :)
        Josh

On Fri, 2006-12-15 at 01:36 -0800, Ken Restivo wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> I've run into the problem outlined in 
> http://lists.nongnu.org/archive/html/fluid-dev/2006-11/msg00041.html and I've 
> bought 2GB RAM for my machine as a "solution" for now. Yay, it works.
> 
> I have a separate, unrelated question too. I'm running 2.6.17 Debian Etch, 
> with realtime-lsm module, 1000Hz timer, and desktop preemption. And, running 
> fluidsynth from the command line or from qsynth gives me endless reams of 
> ALSA overrun error messages:
> 
>       "fluidsynth: warning: ALSA sequencer buffer overrun, lost events"
> 
> Of course, I'm also getting xruns on the audio side too, periodically. At 
> some point I'll try desktop preemption patches and see how it goes. But I 
> have to wonder if there's something else afoot here.
> 
> Thanks!
> 
> - -ken
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1 (GNU/Linux)
> 
> iD8DBQFFgmyke8HF+6xeOIcRAsbZAJ94Jiec+tKEKPhLHBo+BgxOOmcnHgCfUeB9
> Y8mLDpvVSsYuIgnv8B/DMFg=
> =apb8
> -----END PGP SIGNATURE-----
> 
> 
> _______________________________________________
> fluid-dev mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/fluid-dev
>