[linux-audio-user] Tracking down overruns

[Jan Depner]

If you're not already doing it, use reiserfs or xfs.  From your pictures
it looks like you might be using ext3.  There are a few processes that
should be dead before you start doing realtime.  I use a script with the
following:

killall -9 autorun 2>/dev/null
killall -9 artsd 2>/dev/null
killall -9 jackd 2>/dev/null
rm -rf /tmp/jack*
killall -9 /usr/lib/ardour/ardourx 2>/dev/null
killall -9 oafd 2>/dev/null
killall -9 xbiff 2>/dev/null
killall -9 envy24control 2>/dev/null
killall -9 /usr/bin/aplay 2>/dev/null

I'm using almost the same hardware setup as you and I hardly ever get
overruns (1700+ and ST Audio DSP2000 C-Port).  One thing that did give
me problems was using IO_support set to 3.  The sync seemed to give me
problems.  I set both of mine to 1 (no sync).  I also kicked up to 512
MB of memory from 256 and that helped some too.  You can see what I've
done to my system at
http://myweb.cableone.net/eviltwin69/ALSA_JACK_ARDOUR.html


Jan

  

On Sat, 2003-09-13 at 15:16, Benji Flaming wrote:
> Greetings all!  I am attempting to initiate a long unattended 2-channel
> audio recording - perhaps lasting for an hour or two - but I am prevented
> from doing so by periodic overruns which occur at regular intervals.  This
> interval changes according to my system configuration.  Currently, I
> consistently get an overrun about 20 minutes after I start recording.
> 
> My system is an AMD Athlon XP 1700+ with 256mb DDR266 memory.  Video card is
> an ATI Radeon 7500/64mb.  Sound card is an M-Audio Audiophile 2496.  I'm
> running Slackware 9.0.  My kernel version is 2.6.test5.  Prior to this
> kernel I was using 2.4.21 with low-latency and preemptible kernel patches -
> with exactly the same overrun issue.  I'm using jackd version 0.79.2,
> Ardour/GTK 0.378.0, and libardour 0.686.0 (all built from CVS source a few
> weeks ago.)  I'm starting jackd with the command:
> 
> jackd -v -a -R -d alsa -d ice1712 -r 44100 -p 2048
> 
> I don't want to clog everyone's mailbox with an even bigger message, so I've
> posted a bit of additional info online.
> 
> Results from running latancytest0.42-png:
> http://www.comevisit.com/NorthernSunrise/latency/3x256.html
> http://www.comevisit.com/NorthernSunrise/latency/test2/3x256.html
> 
> My .config file from my most recent 2.6.test5 kernel build:
> http://www.comevisit.com/NorthernSunrise/latency/.config
> 
> My jack, and ardour version numbers, and hdparm info for hda and hdb:
> http://www.comevisit.com/NorthernSunrise/latency/configuration
> 
> I can see from the latency tests that *something* is causing spikes,
> particularly during disk write, but I'm not sure how to determine precisely
> what it is.  Is there any way to obtain a snapshot of processor usage at the
> exact moment of an overrun?  Any help or advice (or reminders about what
> info I've neglected to give) would be very much appreciated!
> 
> |)
> |)enji
> 
> Benjamin Flaming
> --------------------
> "The trouble with computers, of course, is that they're very sophisticated
> idiots."
>