[ardour-dev] Ancient history? - Problem identified

Mark Knecht markknecht at gmail.com
Thu Nov 11 09:56:37 PST 2004


On Thu, 11 Nov 2004 17:38:49 +0000, cv223 at comcast.net <cv223 at comcast.net> wrote:
>  
> A few days ago I posted the following message about tracking down xruns
> using Ardour to record to a SCSI Raid0.  At the end of the message, I asked
> whether anyone had any ideas about why commenting out a single line in
> streamview.cc (which at the very least eliminates updating the "progress"
> rectangles during recording) completely eliminates my xrun problem, in a
> rather dramatic fashion (dozens of xruns per min to none).  Since I got a
> goose egg in response, I thought I would try again because I've run out of
> things to test and don't know where to look next.  
>   
> Basically, how can display updates in Ardour during recording cause xruns
> with writes to the hard drives? 
>   
> Since this problem doesn't seem to be common, it seems that it has more to
> do with my system/hardware configuration than with Ardour/jack.  But what, I
> don't know. 
>   
> Any suggestions would be greatly appreciated.  Thanks for your time. 
>   
> Joel 

Joel,
   It's a strange one. I'm sure that if it was a common problem yuo
would have gottem more response, or possibly even that Jesse would
have never committed the code in the first place as he would have seen
the issue.

   So, my first, somewhat obvious thought, is that drawing pink boxes
is a process of moving a number of pixels to the screen. I presume you
are using a more modern AGP based archtecture? If you are running an
older PCI graphics adapter then you're moving those pixels across the
same bus that your audio is moving across to go to and from the disk
which would probably cause some problems.

   Possibly you need to look at your kernel and your X configuration.
Are you using accelleration? Are you using write combining or anything
else interesting in the kernel?

   Since I don't program and know nothing about GTK I can only ask
what version are you using? Have you recompiled it recently in case
this is some mismatch between the code on your system and something
that could have changed like glibc?

   Even though you tried what appears to be a similar kernel to the
one I'm using did you set up the kernel to make your sound card
interrupt operate non-threaded? That was a useful thing for me to do.

   Sorry that I cannot offer any real answers. I suspect that few
would be able to say that 'X caused your problem' anyway.

Good luck,
Mark



More information about the Ardour-Dev mailing list