On Thu, 25 Jun 2009, Sebastian Moors wrote:
> Oh, no problem. I think the zombie stuff is more important, so qtxml can
> wait :)
> I'm going to test the zombie-branche tomorrow on an older laptop, hope
> that i can help
> you with that testing somehow!
> - Sebastian
Thanks! I've found these to be effective when chasing this bug:
* Use 'nice'... both ways.
$ nice -n 20 ./hydrogen
$ nice -n -20 ./hydrogen
This changes the way that the kernel schedules thread
* Get your computer in a state where it's having to
do a little memory swapping.
* I only have a single PIII on all my computers, so
I'm not sure how much multiprocessors or hyper-
threading makes a difference.
* Make sure you're starting jackd with -Xseq.