[PlanetCCRMA] jackd xruns problem: solved? (FC2 2.6.7rc4 O7kernel)

[Fernando Pablo Lopez-Lezcano]

On Sun, 2004-08-15 at 08:02, Arthur Sosa wrote:
> On 14 Aug 2004 18:14:02 -0700
> Fernando Pablo Lopez-Lezcano <nando@ccrma.Stanford.EDU> wrote:
> > I have the latest  and greatest (as of yesterday, 2.6.8 final was 
> > released in between :-)
> 
> And now also 2.6.8.1 and Ingo's O8 patch!
> Probably they are everybody in a great hurry for the summer holidays...
> You are right when you say gurus are hard working now.
> 
> >  I have just two _experimental_ packages - you have been warned, again -
> >  that are available:
> >    kernel#2.6.7-0.1.rc4O7.ll.rhfc2.ccrma
> >    kernel-module-alsa-2.6.7-0.1.rc4O7.ll.rhfc2.ccrma
> 
> I can't see great differences between this and the previous kernel.
> I must admit it seems to me my system has a little better behaviour,
> and I get less xruns than before when I launch
> # jackd -Rv -d alsa -n 3 -r 44100,  namely 2 or 3 per second,
> but I still get those without other apps running!

Hmmm, which means that you are running with the default "-p 1024". That
is pretty bad behavior. I'm able to run "-p128 -n2" with no xruns
(except for the video card, known problem). 

> I tested one more time jack with CK patches, and I don't get xruns
> even under heavy load, when my CPU is sweating rendering a shot
> with Cinelerra and playing music with two players and running Firefox.
> 
> Wait, I tried another time with the new Planet kernel:
> # jackd --realtime -Rv -d alsa -n 5 -r 44100
> and I get only a xrun per second. But I don't think it's
> a quite good latency.

A few questions: what hardware are you running on? Video card and
chipset? Did you add the "tunings" that I suggested in my email? What do
you see in:
  cat /proc/sys/kernel/voluntary_preemption
  cat /proc/sys/kernel/kernel_preemption

How are your interrupts being used?
  cat /proc/interrupts

So many questions :-)
-- Fernando